Regulatory Impact Analysis,
 Oxides of Nitrogen Pollutant Specific Study
          " "          and •
       Summary and Analysis  of Comments
   Control of Air Pollution from New Motor
   Vehicles and New Motor Vehicle Engines:
    Gaseous Emission Regulations for 1987
 . and Later Model Year Light-Duty Vehicles,
      and for 1988 and Later Model Year
  Light-Duty Trucks and Heavy-Duty Engines;
  Particulate Emission Regulations for 1988
and  Later Model  Year  Heavy-Duty Diesel  Engines
                  March 1985
       Environmental Protection Agency
         Office of Air and Radiation
           Office  of  Mobile Sources

         Regulatory Impact Analysis,
 Oxides of Nitrogen Pollutant Specific Study
       Summary and Analysis of Comments
   Control of Air Pollution  from New Motor
   Vehicles and New Motor Vehicle Engines:
    Gaseous Emission Regulations for  1987
  and Later Model Year Light-Duty Vehicles,
      and for 1988 and Later Model  Year
  Light-Duty Trucks and Heavy-Duty  Engines;
  Particulate Emission Regulations  for  1988
and Later Model  Year Heavy-Duty Diesel Engines
                  March 1985
        Environmental  Protection Agency
          Office  of  Air  and  Radiation
           Office of Mobile Sources

                       TABLE OF CONTENTS


1.    Introduction	    1-1

     I.     Organization	    1-1

     II.    Background of  the Regulations	    1-1

           A.     Clean Air Act Requirements	    1-1

           B.     Regulatory History	    1-2

     III.   Description of the Action	    1-3

      - -••  A.     New Emissions Standards	    1-3

           B.     Particulate and Nox Averaging	- 1-4

 - - - - - -— c.  -   New Allowable Maintenance Regulations  .    1-4

           D.     Test Procedure Revisions	    1-5

     IV.    List of Commenters	-	    1-5

2.   Technological Feasibility  	    2-1

     I.     Introduction	    .2-1

     II.    Light-Duty Trucks (LDTs) 	    2-1

           A.     Synopsis of NPRM Analysis	". .    2-1

           B.     Summary and Analysis of Comments  ....    2-3

           C.     Conclusions	    2-20

     III.  Heavy Duty Gasoline Engines  (HDGEs)  	   2-21

           A.     Synopsis of NPRM Analysis	   2-21

           B.     Summary and Analysis of Comments  ....   2-23

           C.     Conclusions	   2-29

     IV.    Heavy Duty Diesel Engines (HDDEs)   	   2-30

           A.     Synopsis of NPRM Analysis	   2-30

           B.     Summary and Analysis of Comments  ....   2-34

           C.     Conclusions	   2-70

                   TABLE OF  CONTENTS (confd)


3.   Economic Impact	    3-1

     I.     Synopsis of NPRM  Analysis	  .  .  .    3-1

           A.    Cost to Manufacturers	    3-2

           B.    Cost to Users  .  .  .  .-	  .    3-3

     II.   Summary and Analysis of Comments	    3-4

           A.    LOT NOx	'	    3-4

           B.    HDGE NOx	    3-20

           C.    HDDE NOx and Particulate Standards  .  .  .    3-32

           D.    Socioeconomic Impacts •	    3-99

4.   NOx and Particulate Environmental Impact 	    4-1

     I.     Overview of NPRM Analyses	    4-1

           A.    Oxides of Nitrogen (NOx)	    4-1

           B.    Particulate Matter 	    4-2

     II.   Summary and Analysis of Comments on NPRM
           Environmental Impact and Air Quality
           Projections	    4-5

           A.    Factors Common to Both Analyses  ....    4—5

           B.    Factors Specific to NOx	    4-14

           C.    Factors Specific to Diesel Particulate .    4-19

     III.  Emissions/Air Quality Projections  	    4-23

           A.    NOx Analysis	    4-23

           B.    Diesel Particulate Analysis  	    4-37

5.   Cost Effectiveness	    5-1

     I.  Overview of NPRM Analysis	    5-1

     II.   Summary and Analysis of Comments	    5-2

     III.  Updated Cost Effectiveness Analysis  	    5-4


                   TABLE OF CONTENTS  (confd)


           A.    Changes in Analysis	    5-4

           B.    Results of Updated Analysis   	    5-5

           C.    Comparison to Other  Control  Strategies  .    5-11

           D.    Conclusion	    5-16

 6.   Alternative Actions   	    6-1

	I.   .-Introduction	    6-1

     II.   Alternative Light-Duty Truck  (LOT)  NOx
           Standards	    6-1

     III.  Alternative Heavy-Duty Engine (HDE)  NOx
           Standards	    6-4

     IV.   Alternative Heavy-Duty Diesel Engine (HDDE)
           Particulate Standards   	    6-6

     Appendix A - Summary  and Analysis of Comments
     on  the  Proposed Particulate Test Procedure for
     Heavy-Duty Diesel Engines   	    A-l

     I.    Recommendations Accepted by EPA	    A-'l

     II.   Recommendations Not Accepted  by EPA	    A-8

     III.  Issues Raised by EPA  in NPRM	    A-21

     Tt/    m-Vjojr  Tcciiae                  .........    A —7.1

                           CHAPTER 1


I.   Organization

     As  required  by Executive  Order  12291,  this  document  has
been  prepared  to   summarize   the  results  of   all  analyses
conducted  in support  of  the  final rule  for  gaseous  emission
regulations  for 1988 and  later model  year  light-duty vehicles,
light-duty  trucks,  and heavy-duty  engines  and  for particulate
emission  regulations  for  1988  and  later model  year heavy-duty
diesel  engines.   In  addition,  this   document  also  provides  a
summary  and  analysis   of  most  of the  comments   received  in
response  to  the  Notice  of Proposed  Rulemaking  (49  FR  40258
October  15,  1984).  Included  here  is a  consideration of  the
technological   feasibility,   economic  impact,   environmental
effects  and cost  effectiveness of  the standards  along  with the
development  of  data  on  the  impacts  of  several  regulatory
alternatives.   The  remaining   issues  raised  by  commenters  to
this  rulemaking are reviewed and responded  to  in the preamble.
These   include   the   proposed   averaging   program,  allowable
maintenance  provisions  and  high altitude  standards.  The oxides
of  nitrogen  (NOx)  environmental impact  analysis  contained in
this  document also serves  as  the  NOx pollutant-specific  study
required by  Section 202(a)(3)(E) of the Clean Air Act.

     The  material  presented  in  this  document  deals primarily
with  those  areas  of  the  draft Regulatory  Impact  Analysis-[l3
which  were  the  subject of  public ^ comment.  Areas  of  analysis
which  were  not  commented  upon  are   repeated  here  only  where
needed  to aid the understanding of material being revised.  The
draft  analysis  is therefore incorporated into  this document by
reference  for  treatment of  topics not specifically  re-addressed

II.  Background of  the  Regulations

     A.     Clean Air Act  Requirements   	-	

     The  Clean  Air Act Amendments  of  1977  created a  statutory
heavy-duty   vehicle  (HDV)   class   and  established   mandatory
emissions  reductions for  that  class.  Under the language of the
amendments,  all  vehicles  over 6,000  Ibs  gross  vehicle  weight
 (GVW)  were defined  as  "heavy duty"  and were required to achieve
a  75  percent  reduction  in  NOx  emissions  from  uncontrolled
levels,  effective with  the  1985 model  year.

     The  Act made no  specific  provisions  for light-duty trucks
 (LDTs),  which at  that  time  only  encompassed LDTs between 0 and
6,000  Ibs GVW  (light  LDTs).   These LDTs were  regulated  by EPA
as  a separate  class  under  the  general authority  of the Clean
Air  Act.  Beginning with the  1979  model  year,  EPA  expanded its
standards  for the LDT  class to 8,500  Ibs GVW,  thus  encompassing
those  heavy LDTs  (6,001  to  8,500  Ibs  GVW)  which are subject  to
the heavy-duty vehicle  provisions mentioned  above.


     The  Act also  authorizes  the Administrator  to temporarily
establish  revised NOx  standards  for heavy-duty  engines  if the
statutory  standards cannot be achieved  without increasing cost
or  decreasing fuel  economy to  an  "excessive  and  unreasonable
degree."T2]   The new  heavy-duty engine  NOx  standards  in this
document  are being  promulgated  under  these  provisions  of the

     The  amendments of  1977  also require  the  "greatest degree
of  Cparticulate]  emissions  reduction  achievable,"  given  the
availability   of  control  technology   and  considering  cost,
leadtime and  energy impacts.[3]   These reductions were to begin
in  the  1981 model  year.   Although  not  specifically  limited as
to  applicability in the  language  of the Amendments of 1977, it
was  recognized   that  the  requirement  was  aimed  at  diesel
engines.    The   heavy-duty  diesel  engine  (HDDE)   particulate
standards  in this rulemaking are based on this  authority.

     B.     Regulatory  History

     The  first  NOx standards antedated  the amendments of  1977*.
Prior  to  the 1975  model  year,  LDTs  complied with  the 3.0 g/mi
NOx  standard that  had been  established two years earlier for
LDVs.   With the  splitting off  of  the  LDT class  for  the 1975
model  year, LDTs  were required to  meet a NOx  standard of 3.1-
g/mi,  comparable in stringency to the LDV  standard.  Heavy-duty
engines  (HDEs)  had no separate  NOx standard  until  the 1985
model  year, however,  there have been combined  hydrocarbon  (HC)
+ NOx standards  in  place  for HDEs  since  the 1974  model  year.

     The  current  NOx   standard  for   1979  and  later  model year
LDTs  is  2.3 g/mi,  comparable  in  stringency  to  the  2.0 g/mi
standard  established for LDVs of that year.  Beginning  in  1979,
the  LDT class  was  expanded  to  include  vehicles between  6,001
and  8,500  Ibs GVW.   The current NOx  standard  for  HDEs is  10.7
g/BHP-hr,  established  originally  for the  1984 model year, but
later made optional until the 1985 model year.

     Turning  now to more  recent  actions, an Advanced  Notice of
Proposed  Rulemaking (ANPIW) was  promulgated for  LDT and HDE NOx
emissions  in January  of  1981  (46 FR 5838).   Standards of 1..2
grams  per  mile  for   LDTs and   4.0 g/BHP-hr   for  HDEs  were
suggested   effective   for  the   1985  and  1986   model   years,
respectively.    These  standards  did   not  correspond  to  the
statutory   75  percent  reduction   as   noted   above,   but  were
proposed  because  they  were comparable in  stringency  to the
existing   1.0  g/mi  LDV NOx standard in  the  case of  LDTs and
because  they represented  what  EPA  believed  at that  time to be
the lowest practicable standard  given the available  technology
in  the case of HDEs.


      The  first  diesel particulate  standards  were  established
 for  LDVs  and  LDTs,  effective  beginning with  the  1981  model
 year.   A standard  of 0.60  g/mi  was  established  for both  LDVs
 and   LDTs,  representing  an  achievable  level  for  the  (then)
 available  technology.  More stringent  standards  (at  0.26  for
 LDTs,   and  0.20   for LDVs)  were   also   promulgated  effective
 beginning  with the 1985  model year, but  these have been delayed
 and  will  now  be effective for  the  1987  model year (49  FR 3010,
 January 24,  1984).   For HDDEs,  a  Notice of  Proposed Rulemaking
 (NPIW)  was  published  in  January,   1981  (46  FR  1910)  which-
 proposed  a standard  of  0.25 g/BHP-hr  for 1986 and  later model

      Because  of the  related technical  issues  that  were  raised
 during  the  comment  periods  for  both  the  NOx ANPRM  and  the
-particulate   NPR4   and  the  interrelationship  between  NOx  and
 particulate  emissions, EPA  decided  to issue a  combined NPRM  to
 address these  issues  to  insure that manufacturers could direct
 their efforts  at  meeting  a unified  set  of  emission standards.
 The  Notice of  Proposed  Ruleraaking  was published on  October 15,
 1984  (49  FR   40258).   This  final   rule,  preceeded  by  public
 hearings  and  a public comment  period, completes the rulemaking

 Ill. Description of the  Action

      A.    New Emissions  Standards         -            '  ,

      This  rulemaking contains  new  low-altitude  NOx  standards
 for  LDTs  and  HDEs,  new  low-altitude particulate  standards for
 HDDEs  and new high-altitude   idle  Co,  NOx  and  particulars
 standards  for  LDTs.   For 1988  and  later  model years,  the NOx
 standard   for  LDTs  is  1.2  g/mi  for  LDTs up  to  and  including
 3,750  Ibs  loaded-vehicle  weight.    The  standard  for  1988 and
 later model year  LDTs over  the above weight limit is  1.7 g/mi.
 A staged NOx standard is established  for HDEs  to allow leadtirae
 for   further   development   of  control   technology.    The  NOx
 standard   for   1988-90   model  year   HDEs  is  6.0  g/BHP-hr,
 representing   a level  that  is  achievable  given  the  available
 leadtime   for  engines currently   in  production,  with   a  more
 stringent  standard of 5.0 g/BHP-hr  effective for 1991 and  later
 model year engines.

      A three-phased  particulate  standard   is  established  for
 HDDEs.   Model  year  1988-90 HDDEs will  meet a  standard of 0.60
 g/BHP-hr.   For 1991-93  model  years,  urban bus  engines  will
 comply with  a standard  of  0.10 g/BHP-hr,  while  the  remaining
 HDDEs  will  meet  a  standard of  0.25 g/BHP-hr.   Both  of  these
 1991 standards will  likely  require  the use  of  trap oxidizers on
 a majority of  applications.  This will be followed by the  third
 phase,  when  all  1994  and  later  model  year HDDEs  will comply
 with the 0.10 g/BHP-hr standard.


     Finally,   certain   new  high   altitude   standards   are
established  for  light-duty trucks.   NOx standards  equal  to the
1.2 g/mi  and  1.7  g/mi low-altitude standards  are established,
along with  an idle CO standard  of  0.50 percent of exhaust gas
flow  at  idle  (gasoline-fueled  light-duty trucks  only)  and  a
particulate  standard  of   0.26  g/rai  (diesel  light-duty  trucks

     B.     Particulate and NOx Averaging

     With   this   rulemaking,  particulate  averaging   will  be
afforded  to manufacturers of  1991  and later model year  HDDEs.
However,  they  will not be allowed  to  average  HDDEs  with LDDTs
or LDDVs  if the  manufacturer' s product line also includes these
vehicle  types.   Similarly,  averaging  California   engines  with
engines  intended for  sale  in  non-California  areas will  not be
permitted,  although  averaging  within  each of  these  areas is
allowed.   Urban  buses will  be  excluded  from  the particulate
averaging  program  to  insure  the  maximum reduction   in  urban
particulate  emissions.  Because  HDDE standards  are expressed in
mass per  unit of  work (g/BHP-hr) .rather than mass per unit of
distance  travelled  (g/mi)  and because  HDDEs  are  divided   into
subclasses  with  widely varying  useful  life  periods,  averaging
will  be  limited  to  within each  of  the existing  subclasses
(light-,  medium-,  and heavy-heavy duty) and  the calculation of •
average  particulate  emissions must  include  weighting factors
for brake horsepower  as well as for production volume.

     NOx  averaging  has  been  established  for  1991 and   later
model year  HDEs and  is   similar  to the  particulate  averaging
program,  with   the   following  exceptions.   The  NOx  averaging
program  is  restricted by fuel  type,   with  gasoline-fueled and
diesel   engines    complying   with  the  standard   by  separate
averages.   For  HDDEs, the  averaging  is  restricted by  engine
subclass  (light,  medium,  and heavy);  however,  gasoline-fueled
HDEs have no such restriction.  Also,  urban buses  excluded  from
particulate averaging  may  be  included  in  the  NOx  averaging
program  for  all  HDEs.

     Finally,    NOx    averaging   for    light-duty   trucks   is
established beginning in  1988.    This   program   is  patterned
closely  after those  established  for heavy-duty engines and the
existing  light-duty  diesel  averaging  program.   Further details
for  both  the  NOx   and  particulate  averaging   programs  are
outlined   in  the   preamble  and   included    in   the   revised

     C.     New Allowable  Maintenance Regulations

     The  allowable  maintenance  provisions  proposed have   been
retained   largely  unchanged.    The  concept  of  emission- and
non-emission-related   maintenance has   been  extended  from   LDTs


and HDEs to  encompass  LDVs  as  well.   Maintenance intervals have
been changed, including  revisions  to the proposed intervals, as
outlined  in  the  preamble.   Manufacturers  will  be  required to
demonstrate  the  likelihood  of  in-use  performance  for  certain
critical emission-related maintenance.

     D.    Test Procedure Revisions

     The heavy-duty engine  test  procedures have  been revised to
incorporate particulate  test procedures.   These  include changes
in  response  to  comments  along with  other  minor  corrections, as
outlined in the preamble to the final rule.

IV.  List of Commenters

     The    following     individuals,    organizations,    public
authorities  and  manufacturers  submitted  written  comment  in
response  to  the NPRM  (49  FR  40258).   This  list  contains only
those  comments  received  by January  4,  1985.   Comments received
after  that  date,  although  not  specifically identified  here,
have  also been  incorporated  fully  into  EPA1s   analyses  along
with those listed.

     1.    Adair, Holiday,  Akron, OH
     2.    American Automobile Association
     3.    American Honda Motor Company
   . 4.    American Lung Association
     5.    American Lung Association, of Berks County  (PA)
     6.    American   Lung   Association   of    Deleware/Chester
           Counties (PA)
     7.    American Lung Association of Florida
     8.    American Lung Association of New Jersey
     9.    American Lung Association of Western Missouri
     10.   American Motors Corporation  (AMC)
     11.   American Public Transit Association
     12.   Arent, Fox, Kinter, Plotkin & Kahn for MEMA
     13.   Arizona Lung  Assoc.
     14.   Audubon Society of Ohio
     15.   Automobile Importers of America
     16.   Bass, Jean, Ross, CA
     17.   Baughman, Jon, Bedford Hgts., OH
     18.   Baumgarten, Sara, Bridgewater, MA
     19.   Bergen County (NJ) Audubon Society
     20.   Bickford, Isabel, Williamsville, NY
     21.   Biesterfeld,  Cathy, Homewood, IL
     22.   Bradman, Asa, Ross, CA
     23.   Brenner, Jeff, New Brunswick, NJ
     24.   Brown, Bruce  and Sharon,  Chicago,  IL
     25.   Brown, Paul M., Sun Olympiad '80
     26.   Burchard, Ann, Robert and  Rachel,  Catonsville, MD
     27.   California Air Resources Board
     28.   California Dept of Justice


29.   Callan, Ida, Vienna, OH
30.   Cape Henry Audubon Society
31.   Capital District Transportation Authority
32.   Caterpillar Tractor Company
33.   Chemel, Bonnie, Evans City, PA
34.   Chicago Transit Authority
35.   Chrysler Corporation
36.   Ciak, Josephine, North Arlington, NJ
37.   Clark County (NV) Health District
38.   Coalition for Clean Air
39.   Coalition for the Environment
40.   Colorado Department of Health
41.   Connaughton, Ruth K.
42.   Cummins Engine Company
43.   Delaware Valley Citizens Council for Clean Air
44.   Delello, Michael, Saranac Lake, NY
45.   Dillon, Mary, Elma, NY
46.   Dolinka, Marvin & Toby, Grand Rapids, MI
47.   East Michigan Environmental Action Council
48.   El Paso Clean Air Coalition
49.   Environmental Alternatives, Inc.
50.   Faulconer, Mrs. James H., Strasburg, VA
51.   Fisher, C. Donald, Muncy, PA
52.   Ford Motor Company
53.   Fox, Warren, Linwood, NJ
54.   Gardiner, Jeffrey, Schenectady, NY
55.   General Motors Corporation (GM)
56.   Geymer, Christine, Oak Park, IL
57.   Gordon, Robin, Great Neck, NY
58.   Greater Cleveland Regional Transit Authority
59.   Grenfo, Louise, Crossville, TN
60.   Group  Against Smog and Pollution
6L.   Hamilton, James, Cleveland, OH
62.   Hawarth, Terrie, E. Grand Rapids, MI
63.   Holmes, David, Clarion, PA
64.   Humphreys, Betsy, Morgantown, WV
65.   Huser, Bill, So. Sioux City, NE
'66.   International Harvester Company  (IHC)
67.   Isker, C., Buffalo, NY
68.   Iwanik,_Mike, Richmond, VA
69.   Jaguar Cars Inc.
70.   Jenner  &  Block  for  Engine  Manufacturers Association
71.   Joan Katz Productions
72.   Johnson, David,  Pueblo, CA
73.   Johnson, Nina, Boulder, CO
74.   Johnson, Rose Mary, Louisville,  KY
75.   Kemp,  Katherine, Chicago Heights, IL
76.   Kulakowski, Lois, Tucson, AZ
77.   LTV Aerospace and Defense Company
78.   League of Women Voters of the Clemson Area


79.   League of Women Voters of the Pittsburgh Area
80.   Lewis, Nana,  Larkspur, CA
81.   Love, John, Boulder, CO
82.   Mack Trucks Inc.
83.   Mannchen, Brandt, Houston, TX
84.   Mansell, Gerda, Lancaster, NY
85.   Manufacturers of Emission Controls Assoc.
86.   Massachusetts Department of Env. Quality Engineering
87.   Mazda (North America)
88.   McCarty, Donna, Indianapolis, IN
89.   McGuire Clinic
90.   Mercedes-Benz Truck Co.
91.   Meyer, Arthur, Akron, PA
92.   Motor and Equipment Manufacturers Association
93.   Motor Vehicles Manufacturers Association (MVMA)
94.   Mueller, Catherine & Edwin, Buffalo, NY
95.   Mueting, Ann, Plymouth, MN
96.   Murkeloff, Robert, Houston, TX
97.   NJ Transit Bus Operations
98.   Natural Resources Defense Counsel (NRDC)
99.   New Jersey Department of Environmental Protection
100.  New Mexico Environmental Improvement Division
101.  Newberry, William, Grand Rapids, MI
102.  Nissan Research & Development
103.  Oakes, Margaret, Boulder, CO
104.  Oregon Department of Environmental Quality
105.  Osterpard, Elsie, Grand Rapids, MI
106.  Otter Creek Audubon Chapter
107.  PACCAR Inc.
108.  Pettit, Marie, Harrisonburg/ VA
109.  Rhode Island Department of Environmental Management
110.  Richmond  (VA) Audubon Society
111.  Rolls-Royce Motors
112.  Rosche, Olga, South Wales,NY
113,  Ross, G.M., Lowell, MI
115.  Saab-Scania of America
'116.  Schiffirth, Anne & Jim, Pittsburgh, PA
117.  Schoenfeld, Josephine, Grand Island
118.  Sherman,  L. Ann, Schaumburg, IL
119.  Shutter,  S.L.
120.  Simpson,  Robert, Flint, MI
121.  Smith, Bertha, Grand  Rapids, MI
122.  South Coast Air Quality Management  District
123.  Southern  California  Rapid Transit District
124.  St.  Cloud Area Environmental Council
125.  Storgul,  Pauline, Chicago. IL
126.  Tonseth,  Phebe, Cumberland Foreside, ME
127.  Toyota  Technical Center, USA
128.  U.S.  Department of  Energy
129.  U.S.  DOT,  Urban Mass  Transit Administration


130.  VIA Metropolitan Transit
131.  Volkswagen of America
132.  Volvo-North American Car Operations
133.  Volvo White Truck Corporation
134.  Volvo of America Corporation - Bus Division
135.  Washington State Department of Transportation
136.  Wedow, Nancy, Palatine, IL
137.  West Michigan Environmental Action Council
138.  White Lung Association
139.  Willard, Dwight, Albany, CA
140.  Williams, Mark, San Francisco, CA



     1.    "Draft  Regulatory  Impact.  Analysis  and  Oxides  of
Nitrogen Pollutant Specific  Study  Control of Air Pollution From
New  Motor Vehicles  and  New Motor  Vehicle Engines:   Gaseous
Emission  Regulations  for  1987  and Later Model  Year Light-Duty
Vehicles,    Light-Duty    Trucks,   and   -Heavy-Duty   Engines;
Particulate  Emission Regulations for 1987 and  Later Model Year
Heavy-Duty Diesel Engines," U.S. EPA, OAR, OMS, October 1984.

     2.    The Clean Air Act  As  Amended August 1977, Serial No.
95-11, Section 202(a)(3)(B)&(C).

     3.    IBID; Section 202  (a)(3)(iii); p. 102. ,

                            CHAPTER 2


 I.    Introduction

      This  chapter  analyzes  the  technical  feasibility  and  the
 leadtime  requirements of  the  final  1988  and  later model  year
 light-duty  truck (LOT)  standards  for oxides of  nitrogen  (NOx)
 emissions  and  the  final heavy-duty  engine 
                               2-2 .

     It was  concluded from this analysis that  the  1.2  g/mi  NOx
standard  was  feasible  for  all  LDGTs   and  for  lighter  LDDTs
(LDDTiS).   The principal  compliance  means   for  LDGTs would  be
closed loop,  three way catalyst  technology,  while  lighter  LDDTs
would  rely  on  the   application  of  EGR.   For LDDT2s  it  was
concluded  (because of  their  heavier weights and larger  frontal
areas) that  a  1.2 g/mi NOx  standard  would  increase particulate
levels  such  that it would  affect  the  stringency  of the  0.26
g/mi particulate  standard.   Consequently, EPA  then  considered a
1.7  g/mi  NOx  standard   for   LDDT2s.-   As  in  the  case   of
LDDTiS, EGR was projected to provide the means  for  compliance.

     It was  concluded  that  a  1.7  g/mi  standard  appropriately
balanced  the  need  for  NOx   and  particulate   control.    As  a
result, EPA  decided  to propose  a. NOx  standard of  1.7  g/mi  for
LDDT:S.   EPA  also  decided  that  it  was   most  equitable   to
propose the  1.7  g/mi  NOx standard  for LDGT2s.  This was  done
because a  more stringent  NOx  standard  for   heavier  LDGTs  would
encourage  the  purchase of LDDT2s,  which EPA   did  not wish  to
do.  Further,  the loss of NOx  control due to  a   more  lenient
standard  for  LDGTzS  was small  compared to   the  case where  only
diesels were affected.

     Considerations  of  the  effects  on  fuel   economy  of  the
proposed  NOx  standards in  combination  with   the  technologies
expected to  be employed  led  to the conclusion  that  LDGTs  which
were converted to three-way catalyst  technology from oxidation
catalyst  technology  could  experience   up  to  an  8  percent
improvement  in  fuel  economy.    For  those  LDGTs which  already
employed three-way   catalyst  technology,  it was projected  that
some small fuel  economy  loss might  occur.  It  was  forecasted on
a  sales-weighted  basis that roughly  a  2-4   percent improvement
in  fuel economy  would be  associated with the proposed standards
for  LDGTs.  For  LDDTs,  consideration   of  the  effects  of  the
proposed standard  on  fuel  economy led to the conclusion that no
significant fuel economy penalty  would  result  from  the proposed
standards  for  either  LDDT,s  or LDDTzs.   This conclusion  was
based on evaluations of the  differences  in  fuel economy between
LDDT,s  with   and  without  EGR   and  on  the   benefits  associated
with the use of electronic controls on LDDT2s.

     For LDTs  sold at high altitude,  EPA proposed  the  same  NOx
standards' as  were  proposed  for  low  altitude   LDTs  because  NOx
emissions  do  not  tend  to  increase  with  altitude.   An  idle' CO
standard  for  high  altitude  LDTs  equal  to  the   0.50  percent
standard  already  required  for  low  altitude LDTs   was  proposed
because a  90  percent reduction from baseline high altitude idle
CO  levels  resulted  in a  numerical value  of  0.51  which,  when
rounded,  was  equal  to the  low altitude value.  A particulate
standard equal to  that at  low  altitude was also proposed.


     B.    Summary and Analysis of Comments

     1 .    Introduction

     The comments  received concerning  the  feasibility  of  the
proposed LOT  NOx standards  are  summarized  and  analyzed below.
When more than  one  commenter  raised  the  same basic  issue,  the
issue  is  treated  once   in   the  summary  and  analysis  with
identification of the multiple sources of  the  comment.   While
the proposed  standards  are applicable  to  two types  of engines
(gasoline and  diesel)  used to power LDTs  and  are  numerically
different as a  function of the weight of  the LOT, comments will
be  treated  by  issue  with  appropriate consideration  of  these
distinctions  where   necessary.   The  issues contained  in  the
summary  and  analysis  of  comments  which   follows   are,   the
technical feasibility of   the  proposed standards,  the leadtime
required for  compliance,  the  effect of the standards  on  fuel
economy and other minor issues.

     2 .    Technical Feasibility of the Proposed Standards

     Six  commenters   provided   comments   on   the    technical
feasibility  of   the  proposed   standards   with  respect   to
gasoline-fueled  LDTs.   Chrysler,   Ford,  General  Motors,  Nissan
and Toyota  stated  that  the proposed standards  (1.2  g/mi  for
LDGTiS   and   1.7   g/mi   for  LDGT2s>   were   technologically
feasible.                  ' •      -  -      --•-

     VW  disagreed  with   the   technical   feasibility  of   the
proposed standard on  the  grounds  that the allowable maintenance
provisions and  the  full-life  useful  life  requirement  result in
requirements  which  are  beyond  the  capabilities  of  current
in-use  or  reasonably  forseeable   technology.    VW  did  not,
however,  provide any information   in  substantiation  of thsir
statement.    Lacking  substantiating  information  for  the  VW
statement  and  considering  the  position   taken  by   the  other
commenters leads to  the conclusion that the proposed  standards
are  technologically  feasible  for LDGTs using  the technologies
identified in the Draft RIA (three  way  catalyst  and closed loop
fuel control).   This  is confirmed by certification data  for the
1985 model  year.  As shown  in Tables  2-1  to   2-7,  nearly all
LDTs   certified  with  three-way   closed   loop  technology  are
already  in  compliance with  1.2/1.7  standards.   Both  full-life
useful  life  and revised  allowable  maintenance provisions apply
to federally certified LDTs for 1985.
     In  the  case  of  diesel  fueled  LDTs,  three
provided  comments  on  the  technological  feasibility  of   -.;
proposed  standards  (1.2  g/mi   and  1.7  g/mi).   One  cornmer.ter
Ford,   stated    that    the    proposed   NOx   standards    ~et

             1985 49-State  Federal Certification Data
      for LDGTs Equipped with Three-Way Catalyst Technology

Light-Duty Trucks
American Motors

Engine Family
- Class 1


Mean Cert.
NOx Level
. 1.70
Light-Duty Trucks - Class 2
American Motors
Ford - -
E3R = Exhaust gas recirculation.
3CL = Three-way catalyst, closed-loop fuel control,
P4P = Air punp.
OTR = Other.
PLS = Pulse air injection.
3WY = Three-way catalyst, open-loop fuel control.
OXD = Oxidation catalyst.
 El = Engine modification.

                               Table 2-2

                1985 49-State Federal Certification Data
        for LDGTs Equipped Without  Three-Way  Catalyst Technology

Light-Duty Trucks
American Motors

^General Motors

Light-Duty Trucks

General Motors

Engine Family
- Class L
- Class 2




•i.o -


Mean Cert.
NOx Level


     See Table 2-1 for definition of terms.

                              Table 2-

            49-State Federal Certification Data for L985 LDDTs
Light-Duty Trucks
American Motors
General Motors
Grumman Olson
Light-Duty Trucks
Engine Family
- Class 1
FTY2 . 4K6 JFT9
- Class 2


Mean Cert
NOx Level

General Motors     FIG6.2K7ZZ42
     See Table 2-1  for definition of  terms.

                                 •Sable 2-4

                  1985 California Only Certification Data
          . for LDGTs Equipped with Three-V&y Catalyst Technology
Emission Control
Bngine Family
Light-Duty Gasoline Trucks - Class
General Motors
Light-Duty Gasoline Trucks - Class

General Motors
piM 2 5ST2HEAO
Mean Cert.
NOx Level
     See Table 2-1 for definition of terms.

                            •Bible 2-5

        1985 California Cnly Certification  Data  for  LDDTs
M anu factur er
Emission Control
Engine Family Technology*
Mean Cert.
NOx Level
Light-Duty Diesel Trucks - Class 1
Nissan - - -
General Motors
Light-Duty Diesel
General Motors
General Motors
Trucks - Class 2
See Table 2-1 for definition of terms.

                            •Bible 2-6

           L985 50-State Certification Data for LDGTs
             Equipoed wich 3-Vfay Catalyst Technology

Manufacturer Engine Family
' Ford
• Zi inner
Trucks - Class 1
-~ PCR2.2T2HB49
Trucks - Class 2
FfoB2 . 2T3FGAO
, F^12.6T6FXX5
"" FCR5.2T2HB^a

NOx Level
See Table 2-1 for definition of terms.

                            •Bible 2-7

            1985  50-State Certification Data for LDDTs
Engine Family
NOx Level
Light-Duty Trucks - Class 1
Grumman Olson
Bd * '"
Light-Duty Trucks - Class 2
See Table 2-1  for definition of  terms.


technologically  feasible  for  LDDTs.    Nissan  stated  that  the
standard proposed  for  LDDT.s  ( L. 2  g/rr.i)  was  feasible  for some
of its  two-wheel drive vehicles,  while its  other  LDDT,s would
need  new  EGR  systems  to  control  NOx  while  simultaneously
complying  with  the  particulate   standard  of  0.26  g/mi.   CM
stated  that  the  ability  of  the 6.2 liter engine  to comply with
a NOx  standard of 1.7 g/mi  through  the  use  of  electronically
controlled EGR would be  marginal  and  would result in increased
particulate emissions  and a  fuel  economy penalty.   GM provided
a figure  (Figure III-C-1 in the  GM comments)  depicting  the NOx
vs.  particulate engine-out emission characteristics for  the 6.2
liter  engine  in  support  of  its  position   that  particulate
emissions would  greatly  increase  under a 1.7  g/mi  standard and
jeopardize its ability to meet the LOT particulate standard of
0.26 g/mi.

     Since all commenters  concurred  either fully  or  with some
qualification  with  the   technological  achieveability  of  the
proposed  standards  for  LDDTs, analysis  of  the  comments need
focus  only  on  the  qualifying  statements   presented  by  the

     In  the  Draft  RIA  (page  2-22),  EPA   concluded  that  all
Federal  LDDT,s  could  be  brought   into   compliance  with  the
proposed NOx standard  (1.2  g/mi)  through the  use  of  EGR  system
designs  already  being used  on counterpart  LDDTis certified to
the California NOx standard.   The thrust of  the  Nissan  comment
is that  a new EGR  system would  have  to  be designed for  use on
some  of its LDDTts  rather  than  the  transfer of  an existing
EGR   system.    Nissan   is,    therefore,  concurring   with  the
technological  feasibility  of  the  proposed  standard by  the use
of EGR  but is identifying  a  need  for  leadtime to design a new
EGR system (leadtime requirements  are  addressed later).

     In responding to  the comment  from GM,  EPA has plotted  1S85
model  year  certification  data for  the Federal  and California
versions of the  6.2  liter GM engine on the  curve provided by GM
in  its  Figure  III-C-1 (reproduced here as  Figure 2-1).  These
values  are  shown  as  F,,   and F2  (1.90  g/mi  NOx,   0.41  g/ni
particulate  and  1.90  g/mi  NOx,  0.34   g/mi  particulate)  for the
two  federal  vehicles  using  EGR   and  as Ci   and  d  (1.7  g/rni
NOx,  0.32   g/mi   particulate  and  1.3  g/mi  NOx,   0.32  g/'nu
particulate)   for     the    two    California    engines    using
electronically controlled EGR.  As can be  seen from Figure  2-1,
NOx  and  particulate  emissions   actually   being,   achieved  are
substantially  lower  than  the gen-eralized curve  presented by
GM.   In addition,  because  the 6.2 liter  GM  engine  is   already
very  close  to  the particulate standard on  an  engine-out  basis,
EPA  can  see  no  basis  for   GM's  concern  about  meeting  tr.e
particulate  standard.   The application  of  particulate traps co

                                              Figure 2-1
                    NOx -  Particulate Trade -  Off Curve —-  6.2L Light Duty Diesel
      0.70 --
      0.60 - -
      0.40 •-
                                                                 Electronic Modulated EGR
                                        Ni)x — Grams Per Mile


 approximately  percent of  these  engines  would,  with  averaging,
 secure  compliance  with  the 0.26 g/mi  standard.  EPA  concludes
 t'hat  the  NOx/particulate  curve supplied by GM is  not  applicable
 to  current  versions of its  6.2  liter  engine,  and that  GM should
 have  no difficulty  meeting a 1.7  g/mi  standard with  this engine.

      3.     Leadtime Requirements

      Four  commenters  stated   that   the   time  required   for
 implementation  of   the  LOT  NOx  standards  for  gasoline-fueled
 vehicles   exceeds  the   time   available   under   the   proposal.
 Specifically,   the   comments  were  as   follow.   American  Motors
"indicated that 34  months are normally  required for the  change
 in  catalyst  technology   required  by  the  proposed  standards.
 General   Motors  stated  that  106  weeks  were required  to  make
 changes   to LOT  bodies   necessary  to  accommodate  the  larger
 catalysts required  for compliance  with the  proposed  standards.
 Nissan  indicated that 2-1/2  to  3  years  would  be  required  for
 changes  to the vehicle  body necessitated  by  the use of  larger
 catalysts.   Toyota  stated that -additional  time  is needed beyond
 that  available  for  compliance by the 1987 model  year  because of
 the  change  in catalyst   technology  necessary  and  the need to
 establish durability and  reliability characteristics of the  new

      Two  of  the commenters  (GM and  Nissan)  predicated  their
 leadtime  requirements on  the need to change the  floorpan  of  the
 vehicle   body   to   accommodate   larger   catalysts.    The  tasks
 required  for the specified  change  in  the vehicle body  would be
 the redesign of  the  floorpan  to  provide the necessary space and
 the  procurement  of  new  dies  for   the   manufacture   of   the
 redesigned floorpans.  Since  the commenters  indicated  in other
 areas of their comments   that  they  have already  quantified  the
 catalyst  volume  requirements, redesign  of  the   fioorpan  can oe
 assumed   to  start  essentially  at the  publication  date of  the
 final  rule.    The   maximum  time  requirement  which  could  be
 allocated for  the  redesign of  the  floorpan  is  6 to 9 months.
 Following redesign  of  the  floorpan,  the  time  required for  the
 procurement of  new  dies  is between  26 and  36  weeks  (Reference 2
 at  page  7-7)   including  the  time  required for  installation of
 the  new  dies  in  the  presses.   In  total,   these two tasks  are
 expected  to require  between 50  and  72 weeks  or  a maximum of 18
 months.   Starting with a  publication date of  March  15,  1985  for
 the  final   rule   and  ending  with  October  1,  1986  for  the
 introduction date of the  1987 model year LDTs,  defines  a period
 of  18-1/2  months for the  execution  of the  tasks necessary cor
 vehicle  body redesign.   Leadtime requirements  for  the  redesign
 of  LDT  bodies  is,  therefore, not  a  viable  basis  for  clai-ir.g
 that  insufficient  time  is  available  for  the implementation ;f
 the proposed standards on LDGTs.


     Turning  to  the comments provided  by American Motors,  that
34 months is the normal  requirement  for  the  introduction of new
catalyst  technology to  a specified  class  of  vehicle,  the steps
required for this work can be identified  as  follows.   The first
step  would  be  the development  of  an  overall  system design
including the  integration of  the new system  into  the  vehicles.
The subsequent steps would  be:   1)  ordering of modified tooling
for  the  manufacture  of   redesigned  components   which  could
include vehicle  body redesign,   2)  the  construction and testing
of  experimental  systems  at  several  calibrations   to  establish
one  or   more  calibrations   for  use  on  emission  durability
vehicles,  3)  collection   of  emission   durability   data,   4)
collection  of data from  emission   data  vehicles,  and  5) -the
incorporation  of modified  tooling  on   machine lines   for  the
manufacturer of  redesigned components.

     EPA's  timing   requirement  estimates for  the  primary tasks
in  the critical  timing  path for gasoline-fueled  LDTs  are  as

	Task	     Time Requirement

Overall System Design and Vehicle Integration  4-6 months
Develop Durability  Vehicle Calibrations        5-7 months
Generate Emission Durability Data              11-12 months
Develop Final Calibrations                     1-2 months
Run Data Vehicles                              1 month
Complete Certification Process
With EPA  and Add Modified Tooling              1-2 months	
                           TOTAL               23-30 months

     Starting with  a publication date of  March 15, 1985 for the
final  rule  establishing the  NOx standard, approximately 18-1/2
months  is  available prior to the start  of the  1987  model  year
in  October  of  1986.  Since the minimum  time  required to perform
the necessary  tasks is  greater  than the  time  available,  it is
concluded  that  insufficient  time was allowed  by  the  proposal.
The addition  of  12  months to the time  available for performing
the  necessary  tasks  by  delaying   the  effective  date  of  the
standards  to  the 1988  model year  would provide  adequate  time
(30 months) for  the performance of the tasks.

     Two  commenters provided statements  to  the effect  that the
time  necessary for  implementation  of the LOT  NOx standards on
diesel  vehicles  was greater  than that  allowed by the  proposal.
American  Motors  stated   that   it  would  have  to   add   EGR  and
particulate  traps   for simultaneous  compliance with the NOx and
particulate  standards  of  1.2  g/mi  and 0.26  g/mi respectively
for   its   LDDTiS  and    that  the  earliest   possible   date  for
completion of  this  work  was  the 1988 model year.  Nissan stated


 that  the  1989  model year  was  the  earliest  possible date  for
 compliance with the  NOx  and  particulate  standard so as  to  allow
 sufficient  time  for the  devel-opment  of  new emission  control

     While   both   commenters  integrated  compliance  with   the
 proposed  1.2 NOx  standard  for  LDDTiS  and  compliance with  the
 1987 model  year 0.26 particulate  standard  into their comments,
 this  integration  is not'as  important  as could be  inferred  from
 the  comments   since  the   requirements  for  the   particulate
 standard  were  promulgated in January  1984.   Manufacturers  have,
 therefore,  had  ample  time  to  plan  and   to  initiate  system
 development  and  any tooling  requirements  associated with  the
 particulate  standard.   Timing  requirements  attributable to  the
 NOx  standard are,  therefore,  the only  requirements  which  need
 to  be   addressed  in  analyzing  these  comments.   Primary  tasks
 necessary for  the  application  of  EGR  for  the first tine  (AMC
 does not  offer  a  diesel  engine  in their  LDTs in California  in
 1985)  or  the  application  of  a new EGR system design  (Nissan)
'are:   "1)  overall  system  design   which   identifies   exhaust
 manifold  changes   to incorporate a  source  for the  recirculated
 gases  and intake  manifold  changes  to  incorporate  a  point  for
 the  introduction   of  the   recirculated gases,  2)  orders  for
 modified   tooling   for   the  manufacture   of  the   redesigned
 manifolds*,  3)  build  and test  experimental systems  at  several
 calibrations  to establish  one or more  calibrations  for use  on
 emission  durability  vehicles,  4)  collect  emission  durability
 data,  5)  collect  data  from emission  data  vehicles,  and 6) 'add
 modified   tooling   to  machine  lines  for  the  manufacturer  of
 redesigned  intake  and  exhaust  manifolds.   Timing  requirements
 for  tooling, to  manufacturer  EGR valves  and plumbing, do  not
 enter   into  the  overall  timing  considerations  because  these
 parts  can  be  expected  to  be  supplied by existing  facilities
 (either  vendor  or r»3nufscturer  owned).   The  timing requirement
 estimates for  the primary tasks  in  the  critical  timing  path for
 diesel  LDTs  are as  follows:

 	Primary Task	         Time Required

 Overall  System  Design        -      -             3-4 months
 Develop  Durability Vehicle  Calibrations         5-7 months
 Generate  Emission  Durability Data              11-12 months
 Develop  Final Calibrations                      1-2 months
 Run  Data  Vehicles                               1 month
 Complete  Certification  Process
 With EPA  and Add  Modified Tooling .        '     1-2 months	
                            TOTAL               22 - 28 months
      Timing   requirements   for  the  procurement  of  a  tooling
      mpdification do  not  enter  the critical  path  timing  Line
      since  a complete new  machine line  can  be delivered  in  18
      to 20  months.[2]


     Starting with  a  publication  date of March 15, 1985 for the
final rule establishing  the NOx  standard,  approximately 18-1/2
months would  be  available  prior to the  start  of  the  1S87 model
year  in  October  of  1986.    Since   the   time  available  for
performing  the  necessary   tasks  is  less  than  the  minimum
estimate  for  the requirements, it  appears  that  there  is merit
in  the  comments.    The  addition  of  12   months to  the  time
available by  delaying the  effective  date  of  the standards  to
the 1988  model year would  provide adequate time  (30 months) for
the performance of the tasks.

     4.    Fuel Economy Effects of the Proposed Standards

     For  gasoline  fueled LDTs, six  commenters stated  that the
proposed  standards   would  result   in  a   reduction  in  fuel
economy.   These  statements   were  made  by  American  Motors,
Chrysler, Ford,  General  Motors,  Nissan  and Toyota.   Only three
of  the  commenters  (American Motors,  Ford  and General  Motors)
however,  provided   numerical  estimates  of   the   effects  of the
proposed  standards on fuel  economy.   Ford  indicated  that a fuel
economy   penalty  of  between  1  percent  and   1.5  percent  was
expected  for  heavy  LDGTs   at  a NOx  standard  of  1.7 g/mi  and
approximately  a  2.5  percent penalty  for  heavy  LDGTs at  a 1.2
g/mi  NOx  standard.   In  addition,  Ford  stated  that  there would
be penalties in»the areas of driveability and performance.

     GM  stated that its  LDGTs  are exhibiting  up  to  a 6 percent
fuel economy penalty when comparison  is  made between 1985 model
year  Federal  (2.3  g/mi NOx, 120,000  miles)  and  1985 model year
California* LDGTs.   GM continued  its statement  by  saying that
the proposed  NOx standard  of 1.2 g/mi NOx with a useful  life of
120,000 miles is more stringent than  the California standard of
1.0  g/mi  NOx  with  a useful  life  of  50,000  miles.  In addition,
GM  stated  that  while   the application of  three-way-catalyst
technology  can  be  expected  to  improve fuel  economy  under   a
constant  NOx  standard,  it  cannot  be  expected  to  either  improve
fuel  economy  or  to prevent  a penalty  under  a  more stringent

     The  comment by American  Motors  was  similar   to   the  GM
comment  in  that  American Motors stated  that its  1985 model year
     1985  California  NOx  standards  for  LDTs  up  to  3999 .  Ibs
     equivalent  inertia weight  are  0.40  g/mi  for  50,000 miles
     with  optional  standards of  1.0 g/mi  for  50,000  miles  or
     1.0  g/mi for  100,000,  miles.   For  LDTs  between  4000  and
     5999  Ibs equivalent  inertia weight  the  standard  is   1.0
     g/mi  for  50,000  miles  with an  optional  standard  of  1.?
     g/mi  for  100,000  miles.


California LDTs were exhibiting approximately  1  mpg*  lower fuel
economy than  its  1985  model  year  Federal' LDTs and  that  the 1.0
g/mi, 50,000  mile California standard  was  less   stringent than
the proposed 1.2 g/mi,  120,000 mile Federal  standard.

     Since  manufacturers  had  placed   such  emphasis  in  their
comments on the effects of the California NOx  standards  on fuel
economy, EPA  assembled paired fuel economy data for  1985 model
year  Federal   and   California   specification   LDTs   from  its
certification  records.   The  criteria  used  in selecting paired
data  were  that  the  same  engine,  by  manufacturer,  had  been
tested  under   the   same  dynamometer   loading   conditions  in
vehicles  equipped  with   the  same  transmission  specification,
number  of  driven wheels  (2 wheel  drive  and/or   4  wheel drive)
and  N/V  ratio.   This  information  is  shown  in  Table  2-7.   The
information shown in Table  2-7  is subdivided into three groups,
those   gasoline  LDTs   employing   the   same    technology  for
compliance  with  the  Federal  and  California  standards,  those
using -different    technologies   and   diesel  .LDTs.    Because
California  NOx  emission  standards  include  several  options and
the  specific  option  used  was  not  always clearly  defined in the
records, exact  distinction  between the  vehicle  emission  levels
and  the  useful   life  requirement  of  the  standard  was  not
achievable.    Distinction   was   possible,   however,   between
vehicles certified  to  the 0.40  g/mi,  the 1.0  g/mi and  the 1.5
g/mi standards and is shown on Table 2-8.

     EPA's  overall  assessment of  the  California versus Federal
comparisons is  that  they are of  limited use  in making  precise
conclusions about  the  effects  to  be  expected from the  new LOT
standards.  This  is  first  of  all  due to  the  fact that  the
California  standards  are   more   stringent   than  the  Federal
standards, 'resulting  in  somewhat  lower emission  levels  than,
will  the  Federal standards.   Under this situation,  there will
be  a somewhat greater  impact on  fuel  economy  associated with
the  California  levels.    In  addition,  consideration  must  be
given  to  the  fact  that  the  Federal  standards  will  not apply
until  the  1988  model  year.   This means  that  time would  be
available for improvements aimed  at overcoming  any fuel economy
penalties which might  currently exist.   Lastly,   it must be born
in   mind  that  California  standards  apply  to  only   a  small
fraction of any manufacturers'  total  LOT sales.   Therefore, the
manufacturers can be  expected to  adopt the lowest  initial cost
approach  to  compliance  with  a  relatively  small  concern over
fuel  economy  effects.  This  will  not be the case in  the  longer
term, when  the entire LOT fleet is -affected.
     One mpg corresponds to between a 4  percent  arid a 6 percent
     fuel  economy reduction  for  American Motors  when based on
     comparisons to the  highest  "highway" fuel economy estimate
     or to the lowest "city"  fuel economy estimate.

                                        Table 2-8

                      L985 Model "tear Light-Duty Truck Fuel Economy
                           Federal vs. California Paired Data*
Fuel Economy
Gasoline, Same
QVJ *** - -
Toyota ***
Toyota ***


3CL -
Combined MPG %



NOx Emissions


- 0.30
Gasoline, Different Technology
04 - ...:
04 ,. - " '
5.7L(1) '


3CL .

Elec EGR




Data  pairing  requirements:   equal  engine displacement,  transmission,  N/V and  inertia
Number  in  (   )  following engine  size  identifies the  number of  engine  pairs  used  in
calculating the mean  fuel economy  values  shown.
No changes occur because this is a 50 state vehicle.  Use of  the same engine for  Federal
and California versions implies minimal fuel economy impact associated with the  required
low NOx  level.


     Regardless  of the  above caveats,  instructive  conclusions
 can  be  drawn from  the v Ca lif o"r_nia  data.   For  those  vehicles
 already  having  three-way technology on  the  Federal  versions,
 there  is no  clear  pattern of fuel  economy change  between  the
 Federal  and  California  versions.   The  data  show  that in  some
 cases  there are  marked reductions  in  NOx  emissions with  little
 or  no fuel  economy impact.   On the  other  hand, some  vehicles
 exhibit  significant penalties.   However,   the higher  penalties
 are  generally  associated with  vehicles  having  California  NOx
 levels well below those needed  for compliance with the  1.2/1.7
 Federal  standard.   In  any event, it  has already  been  noted that
 vehicles already  equipped with three-way technology are  largely
 already  in  compliance  with  the 1.2/1.7 standards.   Therefore,
 no  changes  will  be  required  of   these  vehicles  and  no  fuel
 economy  impact will occur.

     For those  systems  configured with  oxidation catalysts  on
 the  Federal  version,   the data  of  Table  2-8  confirms  EPA's
-analysis -from  the   Draft   RIA.    All  cases   switching  from
 oxidation catalysts  to  three-way  catalysts,  except  for  some
 certifying  to unnecessarily  low NOx  levels,  show  a  significant
 gain  in  fuel economy.

     Overall,  EPA  draws  the .following  conclusions  regarding
 fuel  economy  effects   on gasoline-fueled  LDTs  of  the new  LDT
 standards.    First,   for   those   vehicles    already   employing
 three-way technology,  compliance or  near-compliance  is  already
 widespread.   Therefore,  no fuel  economy impact will  result from
 the  new  standards.   Second,  for those  vehicles switching  from
 oxidation  catalysts  .to  three-way   systems,  a   significant
 improvement  in  fuel   economy   should  result  from  the  new
'technology   at   the  NOx  levels associated  with   the  Federal
 standard.    In  total,   there  will  probably be  some  small  fuel
 economy   gain  associated  with   the   new  standards.   Since  the
 amount cannot be precisely quantified at  this time,  no specific
 benefit  will be  included in the economic  analyses  of the  new

      In  the case of LDDTs, GM is  the only  commenter  to comment
 on the  fuel economy  effects of the proposed  standards.   The
 comment  provided by GM was  directed to  their 6.2  liter  engine
 and  indicated  that  GM  expected a  fuel  economy  penalty as  a
 result of the 'proposed  1.7 g/mi standard  which would be greater
 than  the 5  percent which they  are  experiencing  under  the 1985
 model  year  California  standard.

      Inspection  of  the NOx  emission levels  for the  California
 6.2  liter engine (Table 2-8)  shows  that the engine  is certified
 to a  1.5  g/mi   standard  in  conjunction  with  the  particulate
 standard of 0.4  g/mi.   At these California  standard  levels,  the
 change,  in  fuel  economy relative to  the Federal  standard  of 2.3
 g/mi,  NOx and 0.60 g/mi  particulate  is 0.1 percent,  i.e. there


 is  essentially  no  difference between  the  fuel  economy values
 developed  for  the  6.2  liter  GM engine  under 1985  model  year
 Federal  and  California  standards.   Since  the proposed Federal
 standard   applicable  to   this   engine   (1.7  g/mi)   is   not
 numerically  as  stringent  as  the California  standard,  EPA sees
 no  basis  for the comment provided by GM.   The conclusion which
 can be drawn from the  information  shown in  Table  2-8 is that
 there  should be no  fuel  economy  effect on  the 6.2  liter GM as  a
 result of  the proposed standard.

     5.    Other Comments

     Other  comments  pertaining   to  the   proposed  LOT  "NOx
 standards  were provided  in the  areas of  the factors to be used
 in     distinguishing     between    LDTts     and    LDT2s,    the
 comparability between the  1.2  g/mi  proposed standard and the
 1.0 g/mi   LDV  standard and  the proposed high  altitude  standards
 for NOx,  idle CO and  particulate.

     Since these comments  are fully addressed in  the  preamble
 to  the final rule, they are  not analyzed  here.  EPA agrees with
 the need  to  correct  the  discriminator   between  LDTts  and
 LDT2s.   However,   none  of  the comments  in  the  other  areas
 substantiate  a  need  for  changes.   Interested   readers 'are
 referred  to   the  preamble  for further details on EPA's  response
 to  these  comments.

     C.    Conclusions

     As  a result of  the  proceeding analyses of  the  comments
-provided  in  response  to the NPRM,  it  is EPA's conclusion  that
 the proposed NOx standards of 1.2  g/mi  for  LOT t s  and  1.7  g/mi
 for LDT2s  are  technologically  feasible  for  the  respective
 groups of  LDTs.   The  technologies expected to   be   used   in
 complying  with  these emission  standard  levels will center  on
 the use  of  three-way  catalyst technology  with closed-loop  fuel
 control  in  the  case  of  gasoline-fueled  LDTs and  on the use  of
 EGR in the case of  diesel-fueled LDTs.

     Analysis of the  comments has led EPA to conclude  that the
 time  required   (leadtime)   for  implementation of  the  necessary
 technologies is greater  than  that  which would be available  with
 an   implementation  date  of  the 1987 model  year.   A  one-year
 delay  in  the  implementation  date  of the  standards  to  the .1988
 model  year would, however,  provide sufficient leadtime.

     Analysis  of  the comments  provided  on the  fuel  economy
 effects of the proposed  standards  has lead EPA to conclude  that
 on   average,  those   LDGTs  which   are   already   equipped   with
 three-way catalyst  technology  will  experience  little or   no
 reduction in fuel  economy  while those LDGTs which  are  converted


 from oxidation  to  three-way  catalyst  technology are expected  to
 experience  increases  in  fuel economy.   In  the  case of  LDDTs,
 the expectation is that there will  be no  measurable change  in
 fuel  economy resulting from  the NOx  standards.   For the  total
 fleet- of LDTs,  the effect  of   the  proposed  standards  on  fuel
 economy  is  expected to be near  zero  but  with the potential  for
 some   improvement   resulting  from   those   LDGTs  which   adopt
 three-way catalyst  technology.

 III. Heavy Duty Gasoline Engines  (HDGEs)

     A.    Synopsis of  NPRM Analysis

     The NPRM  analysis[1]   examined  the   feasibility   of  the
 proposed 1987  6.0  g/BHP-hr  and the 1990  4.0  g/BHP-hr  NOx
 emissions  standards for HDGEs.   The  analyses for each  standard
 began  with  the identification  of  the  appropriate   low-mileage
 target   values.    Current   HDGE  emission  levels   were   then
.discussed  as part  of  the  analysis   of the  1987  standards,  as
 well   as the   effects  of  leadtime   constraints  and available
 emission control   technologies.   The analysis   for  the  1990
 standard considered  the  likelihood  of  new and  more   refined
 emission control  technologies.   A summary  of the NPRM  analysis
               IT.             (  "
      1.    6.0  g/BHP-hr NOx  Standard

     The  factors   considered  in  estimating   the   low-mileage
 emission target were the additive deterioration factor   and  the
 production   variability  factor.   The   additive  deterioration
 factor   (DF)   was   developed   from   1983  model   year   HDGE
 certification   data  and  was found   to  be  zero.   The  NPRM's
 production   variability factor  of  1.2  was  the mean   of  two
 estimates  previously provided by  Ford and  GM in  response  to  an
 earlier  rulemaking.   These   two  factors  were employed  in  the
 following  equation  to  develop  the  low-mileage emission targec
 of 5.0 g/BHP-hr.

                       Emission Standard - Deterioration  Factor
 Low Mileage"Target  «
                               Production  Variability

      The second step in the  analysis  was the identification  of
 the reductions  in  emissions  required to  meet the target level.
 This  was accomplished  by  a  comparison of the low-mileage target
 and the most  up-to-date  information on-  low-mileage  emission
 levels   actually   being  achieved.    The  most  up-to-date  data
 available  were  those  developed  from  prototype  1985 model year
 HDGEs.   As a result of  the  comparison it  was found  that  four  of
 eleven  engine  families could presently  (1985)  comply with  the
 low-mileage   target.   The   average   reduction   necessary  for
 compliance  with  the standard by the  remaining seven  families


was   15   percent.    The   greatest   reduction  necessary   for
compliance  by an  engine  family was  34  percent and  the  lowest
reduction was 7 percent.

     In  considering  leadtime,   the   analysis  noted  that  some
development   testing  had  already been  performed  and  further
development testing would be initiated during  the  course  of the
rulemaking.   Still,   somewhat  less  than the  equivalent  of  two
years  leadtime  was determined  to  be available  for  NOx control
development,  thus,  precluding  the availability  of major  engine
or hardware changes for production.

     The  final  step  in the  analysis was  the  identification- of
technologies  which  could  provide  the  reductions  in   NOx
emissions   necessary   for    compliance.      Three    potential
technologies  were  identified:    ignition   timing  retard,  fuel
enrichment  of the air-fuel  charge  delivered  to the  engine and
EGR.   Ignition  timing  retard  as  the sole method  of achieving
compliance  was  judged to  be  unacceptable since  it  would  result
in  a  relatively  large fuel economy penalty.   Fuel  enrichment
was  also judged  to  be  undesirable   since  it would  negatively
impact compliance  with both  the HC and the CO standards as well
as  causing   a  reduction  in   fuel   economy.    Increased  EGR,
possibly  coupled  with a  small  amount  of  timing  retard, -was"
judged to be  the approach which  would most probably be employed
by  manufacturers  since the  necessary reduction  in NOx could be
achieved with an insignificant effect on fuel economy.

     The  analysis  concluded that,  based   on  information  then
available,  and  considering   the  relatively  modest  reductions
necessary  for only a fraction  of  the fleet  and  based  on the
availability  of  well understood  NOx control  technologies for
gasoline-fueled  engines,  a  1987 NOx standard  of  6.0 g/BHP-hr
was  feasible  for HDGEs.

     2.     4.0 g/BHP-hr NOx  Standard

     The   low-mileage emission  target  for  the  4.0  g/BHP-hr
standard  was  developed using  the same  procedure  as  that  used
for  the   6.0 g/BHP-hr  standard.    The   same   values  for  the
deterioration factor  and  the production  variability factor were
also used because compliance with the 4.0  g/BHP-hr standard was
expected  to be  achievable without  the use  of  reduction catalyst
technology.   The  low-mileage emission target  developed by this
procedure was 3.3  g/BHP-hr.

     The   reductions   from   1985 model  year  prototype  levels
necessary  for  compliance  with  the  4.0  g/BHP-hr  standard were
estimated   once   the   low-mileage   target   level    had  been
identified.   The  .average  reduction necessary for  compliance was
found  to  be  39  percent  with   the  greatest reduction  being 57
percent  and the least reduction  being 3 percent.


     At  the  level of  emission  control required  for compliance
with  a  4.0 g/BHP-hr,  it was  concluded  that  emission  control
technologies beyond  those  required for compliance  with  the 6.0
g/BHP-hr standard  {i.e., standard EGR)  could  be   required  to
avoid significant  performance  and fuel economy  penalties.  The
technologies identified as being the most  probable  for use were
increased  EGR  rates  with   improved  controls  and  "fast-burn"
combustion   chamber  design,  coupled   with  probable   use  of
electronic   control  to  optimize  fuel  metering  and  ignition

   - With  respect  to leadtime,   the  adoption and demonstration
of these control  technologies were considered, at  that  time to
be feasible for 1990,  based on  the fact  that  prototype  engines
were  already   approaching   the  design  target  and  considerable
experience   was  directly transferable  from  work in  light-duty
vehicle and light-duty truck NOx control.

     B.    Summary and Analysis  of Comments     '       "    -----

     The Agency received comments  on  its  NPRM analysis from the
three manufacturers  of heavy-duty gasoline  engines:  Chrysler,
Ford, and  General Motors.    Their  comments on the  6.0 g/BHP-hr
standard are  examined first  followed  by  an analysis of  those
pertaining  to the 4.0 g/BHP-hr standard.

     As  will  be  seen  in  the  next  section on  HDDEs,  the 6.0
g/BHP-hr NOx  standard  will  not  be  feasible  for   HDDEs  until
1988.   Thus,  this implementation  date will  be  assumed here, as
well.   Also,  the  4.0  g/BHP-hr  standard   was  found  not  to  be
feasible for  HDDEs by  1990.  However,  a  5.0  g/BHP-hr standard
appears  to be  feasible  for  1991.    Thus,  this  will  be  the
second-stage NOx standard considered here for HDGEs.

     1.    6.0 g/BHP-hr NOx Standard

     None  of  the  manufacturers  disagreed with  the low-mileage
target   level  of   5.0  g/BHP-hr,   nor   with   the   low-mileage
prototype  data presented  in the NPRM analysis.  With respect to
the availability of control  technology  and leadtime,  two of the
three   manufacturers   were   generally   in  agreement  with  the
conclusions reached  in  the  NPRM  analysis.   In  their submittals,
both Ford  and  Chrysler stated that they could meet  the proposed
6.0 g/BHP-hr  standard;  Ford in  1987  and  Chrysler  in  1988.  GM
stated  that  this   standard  should  be  feasible  for   its  HDG
vehicles  above 14,000  Ibs.  GVW,  "but  would  result  in  a  1.5
percent  fuel  economy  penalty;   GM added  that  for  its  engines
used  in 8,500-14,000 Ib.  GVW vehicles, it did not  believe that
the proposed  standard  was  feasible in conjunction with  the 1987
model   year   1.1/14.4   g/BHP-hr   HC/CO    standards.    As   an
alternative to the 6.0 g/BHP-hr level, GM recommended a HDE NOx


standard  of  8.0  g/BHP-hr.   However,  this  was  due  to  GM's
continued belief  that  catalyst*technology is still not feasible
for  these engines,  and not  on  an  inability to  meet the  NOx
standard, per se.   Thus,  given  the  fact that  the  initial  NOx
standard  is  being  delayed  to   1988   for   HDDEs,   all   three
manufacturers essentially  agree  that the 6.0 g/BHP-hr standard
is feasible for HDGEs.

     With respect  to  the  technology needed to comply  with this
standard, both Ford  and GM disagreed with the NPRM's 'assessment
that  increased  EGR,  possibly coupled  with  a  small  amount  of
timing  retard,  was sufficient and the  most  likely  approach  to
be employed.  According to Ford,  more  than  just  increased "EGR
and  ignition  timing  retard are  required  in order to comply with
the  regulations while  maintaining  the fuel economy,  performance
and   driveability  of  Ford's   heavy-duty  vehicles.    In  its
confidential comments,  Ford listed the  control  techniques it  is
planning  to  incorporate  in  order  to  meet   a  6.0  g/BHP-hr

     GM  also  criticized  the  Agency's  assessment  of  EGR as  a
control  technique  because  of  the  fuel economy penalty resulting
from  increased  EGR.    However, unlike Ford,  GM  did  not  believe
that  alternative  techniques  were available for its  HDGEs.'  GM
supplied  data taken on a  1985  350-4 V8   prototype  engine that
showed  a  1.5  percent fuel penalty  resulting from an increase in
the  EGR  in  order  to  comply  with  the proposed  standard.   Also,
both  recalibration of  the air-fuel ratio and retarded ignition
timing were found to be unacceptable by  GM for the  same basic
reasons  as  identified  in the NPRM.  Chrysler did not comment on
the  technology  needed for its  engines  to  comply  with   a  6.0
g/BHP-hr  standard.

     Neither  Ford  nor  GM presented sufficient justification for
their  projections of  technology  requirements  to allow  them  to
be  objectively  critiqued  here.    However,  an analysis of 1985
Federal  HDGE  certification data  confirms  the conclusion  of  the
NPRM that EGR is  basically  capable  of  providing  the degree of
control  necessary to meet the 6.0  g/BHP-hr  standard  (see Table
2-9).   Two  engines,   a 7.5L Ford and  a  7.4L   GM,  are   already
being   certified   at   NOx  levels   of  4.2  and  4.5  g/BHP-hr,
respectively.   The  only   significant  difference  between these
engines  and those at  higher NOx  levels  appears  to  be increased
EGR  and recalibrated engine  parameters  (i.e., timing, secondary
air  rates, etc.).  Thus, more significant changes  should  not be
required  for most  HDGEs.   As   roughly  one-third  of  all 1985
prototype  HDGEs  were  able to  comply with  a 6.0  g/BHP-hr  NOx
standard  and  another  one-third  were within 25 percent  of  the
standard, these  engines should  require no more than increased
EGR  rates  plus   recaLibration.   However,  as  described  in  the
Draft  RIA,  the  NOx  levels  of   some  of  the  engines  were well

                                       •teble 2-9

                   1985 HDGE Federal Certification Results (g/BHP-hr)

Manufacturer    Displacement    Emission Control   NOx  (DF)     HC (DF)         CO (DF)

Ford                 4.9              EGR-Air      8.49(.0l)    1.82(0.0)     15.65(.00)


Chrysler             5.9L            BGR-Air      7.7K.07)      .65(.00)     18.73(.00)
Displacement Emission Control


- 4.8
NOx (DF)
1.47 (.02)


above  the design  target of  5.0 g/BHP-hr  (i.e.,  more  than  25
percent).  Compliance by these "engines,  which represent roughly
half of  those not  already  in compliance with  the 6.0 g/BHP-hr
standard,  may require  more  significant  modification  to  avoid
impacting  either   HC/CO  emissions  or  fuel  economy.   These
modifications were  among those  identified in the  NPRM analysis
for  the  4.0  g/BHP-hr standard  and  include  modifications to the
combustion  chamber,  the  intake manifold,   the  secondary  air
system,  and   the  camshaft.   As discussed  in Chapter  3,  these
changes may require some retooling,  but,  given  their nature and
the  comments  of   Ford   and  Chrysler,  leadtime  should not  be

     The  certification  levels shown in Table 2-9  are generally
higher  than   those  of  the prototype  engines described  in the
NPRM.  This does  not necessarily  imply that the  levels  of the
prototype  engines  were  not  in  the  end achievable.  The current
NOx  standard  puts  little,  if  any, real  pressure on  HDGE NOx
emissions,   so   the   higher  certification   results  probably
involved  recalibration   to  higher  NOx  levels.  Thus,  this  does
not  negate the potential  to  achieve the  lower  NOx levels with
the  two sets of engine  modifications described above.

     Also  to  be  noted   from  Table  2-9   is   the  positive
relationship  between HC and NOx emissions  (i.e.,  HC emissions
decrease  as NOx  emissions  decrease).   This  is not  to  say that
EGR  decreases HC  emissions,  but that  other engine parameters,
such  as  the  secondary  air injection rate,  can be  adjusted  to
eliminate  any  adverse   effect  of  EGR on  HC  emissions.   This
positive  relationship  is  present  even  at  the  two  lowest NOx
levels of  4.2 and 4.5 g/BHP-hr.

     With  respect  to fuel  economy,  GM  argued for a 1.5 percent
penalty,  while  the other  two manufacturers  did  not comment  on
the  NPRM's projection  of  no  penalty.   GM based its judgment on
testing  of a  single engine with varying  EGR rate.   It was not
clear  from the information presented  if BSFC  was optimized  at
each EGR  rate,  or if EGR was simply increased.   No actual data
nor   engine    calibrations   were   presented.   Thus,   the  GM
projection  cannot   be   evaluated   against  the   other   three
projections  of  no  penalty.  Thus,  the conclusion  of  the NPRM
will be carried forward here, that  of no fuel penalty.

      In  summary,  essentially all  three manufacturers  of   HDGEs
are   in   agreement  with the Agency's  conclusion  that  a 6.0
g/BHP-hr  standard  is feasible for  1988 model  year HDGEs.   This
standard  is  obtainable  for HDGEs within the available  leadti.T.e
constraints,  and   should   result   in   no  undue   fuel   economy,
performance,  or driveability  penalties.


     2.    5.0 g/BHP-hr NOx Standard

     In  comments  on the  proposed 1990 model  year 4.0  g/BHP-hr
 NOx  standard,  the manufacturers  uniformly  termed this  standard
 infeasible.  Chrysler  did not believe that  the  technology which
 will be  available  by  the 1990  model year  will  be  capable  of
 achieving  the  4.0  g/BHP-hr  standard.   Thus, Chrysler  felt  that
 the  Agency did not realistically assess  the prospects  that  the
 necessary  control  technology  could  be  produced  in  time  to
 assure compliance.
     GM  reported  that its  effort to  reduce NOx emissions  from
 HDGEs  used in trucks  above  14,000  Ibs.  GVW from current  levels
 to  the  level required to comply  with  the proposed standard,  HC
.emissions  were doubled  and  fuel  consumption was  increased  by
 about  6  percent.   Thus,   GM  believed that  the  4.0 g/BHP-hr  NOx
 standard  was  not  feasible because  it  would prevent  compliance
 with the  1.9  g/BHP-hr non-catalyst HC standard for  1987 model
 year  heavy  HDGEs;  also,  the  fuel  assumption  penalty   was

     Ford  contended   that  EPA  erred   in  its   technological
 feasibility  assessment of the  control methods  required to  meet
 the  standard.   Ford was   convinced  that  in  order to  reduce  NOx
 emissions  to  the  4.0  g/BHP-hr  level,  a  three-way-catalyst  was
 required.   According  to  Ford,  a   three-way catalyst  is   not
 capable  of  operating  under  the   high-temperature  conditions
 encountered  by   Class   IIB,   III,   or   VI   heavy-duty  trucks.
 Therefore, it  determined that the  4.0  g/BHP-hr  standard was not
 feasible.   Ford  also  questioned EPA's  analysis  of  fast-burn
 technology "as  a   control method;  Ford  believed  that  the  burn
 rates  of  the  fast-burn   cylinder heads  described  in  the  NPRM
 analysis   will    not   be   significantly   different   than   a
 conventional head at  tne high  speed  and  load conditions cf  the
 heavy-duty transient  test  cycle,  thus making  no  allowance  for
 further  EGR  optimization.

     Since  the   4.0   g/BHP-hr   standard  is  no   longer   being
 considered for HDGEs  the above  comments pertaining  to the  4.0
 g/BHP-hr   standard must  be  analyzed  with  respect  to  a   5.0
 g/BHP-hr   level.   However,  little  detailed technical  analysis
 was  provided  by  the  commenters  to  contribute  to   a  detailed
 assessment of  either a 4.0  or  5.0  g/BHP-hr standard.  Thus, the
 analysis  here  will rely  on  the  analysis  performed for  the  NPRM
 and 1985  certification   data.   Further,  an adoption of  a  5.0
 g/BHP-hr   NOx   standard   should.  mitigate    many   of    the
 manufacturers' concerns.

     The  NPRM  analysis stated  that  the  low-mileage target for a
 5.0 g/BHP-hr  NOx  standard   would  be  4.2  g/BHP-hr.    Based  on
 1985-1987  prototype data, that  analysis  also showed two engines


already to  be  below this  level  and the  remainder  requiring  an
average 30  percent reduction  in  NOx  emissions.   Available 1985
Federal certification  data  (Table 2-9) basically  confirm this.
One engine  is  at  the  4.2 g/BHP-hr target, while another is just
slightly  above this  at  4.5  g/BHP-hr;  these  levels  are  being
achieved  essentially  with  EGR and minor  engine recalibration.
The  remaining  1985  engines  require   somewhat  more  than a  30
percent reduction  on  average.   However,  this is not significant
since  the current 10.6  g/BHP-hr  NOx  standard puts  no pressure
on  NOx emissions,  and,  therefore,  there  was no  guarantee that
the low  NOx levels achieved  by prototype engines  would appear
in  certification.    Given   the   fact  that  two   engines   in
production  already  essentially meet the  low-mileage target "and
a third protoype  engine  also met this level over a year ago,  it
is  difficult  to  argue that this level will  not  be  feasible  six
years  hence.   This   is  especially   true  given   the  general
homogeneity of  HDGE technology,  which stands in  stark contrast
to  the   heterogeneous   HDDE  technology.   The    technologies
discussed  in  the NPRM are  applicable to  any HDGE.   Thus,  the
5.0 g/BHP-hr NOx standard should be feasible for HDGEs.

     This  standard will  require  control  technology similar  to
that  required  for the 6.0  g/BHP-hr standard  (i.e., combustion-
chamber modifications, improvements to the intake manifold,'the
secondary  air system  and  the camshaft).   However,  because  a
greater   level  of  NOx  reduction  is  required  to   reach  5.0
g/BHP-hr,  a larger percentage  of  the fleet  will  require these
hardware  modifications  in  addition to increased EGR  rates  and
recalibrations;  burn   rate  improvements,  as described   in  the
NPRM  analysis,  may also be  required as  a  control technology.
Since  roughly 15  percent  of  the  current HDGEs  of  Table  2-9
essentially comply  with  the 5.0 g/BHP-hr  standard without these
hardware  modifications   and   assuming  NOx  averaging   to   be
available,  it  is  estimated  that  roughly  one-third  of  the
remainder  will  be able  to do  so  as  well.   Therefore,  of  the
approximately  85  percent of  the  fleet requiring any  additional
control,  about  two-thirds  will  require  the hardware  changes
described   above,  in  addition  to  increased  EGR  and  engine

     Although   the   5.0   g/BHP-hr   should   be   feasible   via
engine-related  changes  as  detailed above,  this does  not rule
out the   possibility  that  manufacturers  will decided  to apply
three-way  catalyst technology to meet the standard.  Class IIB
and III HDGVs will be  equipped with oxidation catalysts in 1987
to  comply  with the  HC/CO  emission   standards  and  their  LDGT
counterparts   will   likely   be   equipped   with    closed-loop,
three-way  catalyst  technology.  Thus,  the  step  to  three-way
catalyst  may  be  considered  by  some  manufacturers.    However,
such   a   change  -is   not  likely,   since  manufacturers  have
repeatedly  emphasized  to   the  Agency   their   position  that


 significant   questions  of  feasibility   exist  for   three-way
 catalysts  in  the  heavy-duty environment.   It was  for  reasons
 such  as  these,  and  their  associated cost  impacts,  that  EPA
 chose  not to  propose  a  three-way  catalyst  based standard  for
 HDGEs 'in  the proposal.

     If  done,  application  of   three-way  systems  would  involve
 increased  initial  vehicle  cost.   However,   fuel  economy  and
 performance  should  improve beyond  current  levels,  as  indicated
 in  Section  II  above  for  LDGTs.    Otherwise,  no   substantial
 adverse  impact on  fuel economy,  performance  or driveability  is
 expected,  due  to  the  substantial  leadtime   involved  and  the
 hardware  modifications  available.  "Thus,  a  manufacturer  would
 only be  expected to apply three-way  catalysts  if  it  resulted in
 a  net  cost-benefit  improvement with  respect  to its  profits  and
 consumer  satisfaction.

     C.    Conclusions

     The   following  conclusions  result   from  the   preceeding
 analysis   of  the  comments   provided   on  the  technological
 feasibility  of  the   proposed   standards  and   from  the  draft
 regulatory analysis performed  in support  of this  rulemaking.

     A NOx standard of  6.0  g/BHP-hr should be  feasible  for 1988
'model  year  heavy-duty gasoline,^  engines.  Roughly one-third  of
 all HDGEs are already  in compliance  with this standard  without
 any  hardware modifications  from their higher  NOx counterparts.
 One-half  of  the  remainder  will  require only increased  EGR  rates
,and  engine   recalibration   to   comply.    The   other  half   will
 require  hardware modifications  in addition to  increased  EGR  and
 recalibration.   Complying with a  6.0 g/BHP-hr  standard  should
 not  result in undue fuel  economy,  performance,  or  driveability
 penalties  for  HDGEs.

     A  NOx standard of 5.0  g/BHP-hr  should be  feasible for 1991
 model  year heavy-duty gasoline  engines.   Roughly 15  percent  of
 current   HDGEs  are  already  in  compliance with  this  standard
"without   any  hardware  modifications  from  their   higher  NOx
 counterparts.   Assuming  that  NOx  averaging  will be  available,
 roughly  two-thirds  of  the remainder will  require  only increased
 EGR  rates  and  engine  recalibration  to comply.    The  other
 one-third will require minor hardware modifications  in addition
 to  increased EGR and  recalibration.   This  increased application
 of  control technology  should  avoid  any measurable  performance
 or  fuel  economy  penalties  at the 5.0 g/BHP-hr  standard level.


IV.  Heavy-Duty Diesel Engines (HDDEs)

     In  developing  the proposed  emission  standards  for  HDDEs,
the  NPRM  analysis[l]  treated  the  process  in  two  distinct
stages.    In   the   first   stage,    the   focus   was  on   the
identification  of   achievable   emission   levels  for  NOx  and
particulate  emissions  in the near-term  (1987).   In  the  second
stage,  the  focus  was  on   levels  achievable  in the  mid-term
(1990).  The second stage  of  the development  process  included
the  evaluation  of  feasible  engine-out  NOx  and  particulate
emission levels as  well  as  the feasibility  of  trap  technology.
The  identification  of  feasible  engine-out  NOx  and  particulate
levels  are  clearly  related;  consequently,  they  are  discussed
together and the  analysis  of trap  feasibility  and associated
particulate  standard  levels is  treated  separately.   Thus,  the
near-term  NOx  and  particulate  standards  are   examined  first,
followed by  the mid-term  NOx  and non-trap  particulate standards
and then the trap-based particulate standards.

     A.    Synopsis of NPRM Analysis

     1.    Near-Term NOx  and Particulate Standards

     The NPRM draft  regulatory  analysis[l] of the technological
feasibility  of the proposed  1987  NOx  and .particulate standards,
6.0  g/BHP-hr NOx  and 0.60  g/BHP-hr  particulate, consisted  of
five  steps and  is  summarized  as follows.   The first  step  was
the  identification  of NOx  and particulate emission  levels from
current  engines.  These  data were broken  down  by HDDE  subclass
(light  (LHDDE),  medium  (MHDDE),   and  heavy (HHDDE)), because of
the  technological  differences in  engine   designs  between these
subclasses.   NOx   emission  levels   were  obtained  from  both
Federal  and  California   certification   data.   However,   as
particulate  emissions are  not  currently  regulated,  these data
had to be  gathered from a variety of sources.

     The second  step  in  the  analysis was  the  determination of
the  low mileage  emission  targets  and the amount  of  emission
reduction    necessary   for   compliance   with   the   proposed
standards.     The   identification   of  the   target   level  was
performed  according  to   the  same basic   methodology  described
above  for   LDTs  and  HDGEs.   With  respect  to  the amount  of
emission   reduction   required,   HDDEs  were  divided  into  two
groups:  indirect  injection  (IDI)   and  direct   injection  (DI)
engines.   In the  case of IDI engines, (engines  manufactured by
GM -and  IH),  it  was concluded that available transient test data
on the  GM  engine  showed  that it  could already  comply  with  the
proposed   standards.   Steady  state   data  on  the  IH  engine
strongly suggested  that   it  also  could comply.   As  for  the  DI
engines, which  constitute  the  majority  of the  HDDE families,
all  exhibited higher  NOx  and particulate  levels than was  the
case  for  the  IDI  engines.   Substantial  differences  between


 various  DI  engine  configurations  were  identified  (naturally
 aspirated,  turbocharged  or  turbbcharged wifeh aftercooling).

      T,he   third  step   in   the  HDDE   analysis   on  engine-out
 emissions  was  the  identification  of  the  technologies  which
 could provide  the  necessary emission  reduction necessary  for
 compliance  with   the  proposed  standards.    The  analysis  for
 LHDDEs   (roughly   equivalent   to   IDI   engines)  was   fairly
 straightforward,  since the available  data indicated  that  these
 engines  were  already at or below the  1987 standards.   The MHDDE
 and HHDDE  (roughly equivalent  to DI engines)   analysis was  more
 complex   and   began  with  an   estimation  of  short-term  BSFC
 improvements,   since  reductions  in   fuel   burned   translate
.directly to  reductions  in NOx  and particulate  emissions.   The
 analysis then moved on to  an  assessment of  the effectiveness  of
 various  techniques to  directly control  NOx and  particulate
 emissions,  including  injection  timing  retard  and aftercooling.
.On   the  ..basis  of   the   wide   variation  in   engine   design
 configurations  present   and   the  disparity  in  emissions,  "it
 appeared that each manufacturer, for  each of its engines,  could
 adopt multiple emission control  strategies for  compliance  with
 emission standards.

      The fourth  step  was  an assessment  of  the  effect  of  the
 proposed 1987  standards  on HC  emissions  and  fuel  economy.   In
 estimating  the fuel economy  effects  of the proposed  standards,
 EPA considered estimates  provided by  manufacturers as  well  as
 estimates  developed by  the  National  Academy  of  Sciences  (NAS)
'(page 2-76  in Reference 1).   The resulting estimated  effect  on
 fuel economy  was  for up  to  a  two percent  reduction initially
 diminishing to zero by the third year  of  the  proposed standards
-(6.0/0.60) .

      Leadtime was  the final  step.   Since  the emission  control
 strategies  expected to  be used  in complying  with  the proposed
 standards    involved   recalibrations    of   injection   timing,
 modification    of   aftercooling   and/or    the   addition   of
 aftercooling   on  some  engines,  the   leadtime  required  for  the
 implementation of  the  proposed  standards  was considered  to  be
 within the  time proposed for  implementation.

      2.     Mid-Term NOx  and Non-Trap Particulate Standards

      In the NPRM analysis, the assessment of  the feasibility of
 the  1990 NOx and  non-trap particulate standards was  performed
 in three steps and is  reviewed as "follows.  The initial step of
 the  1987   analysis,  the  determination^  of   current  emission
 levels,  did  not  have  to be repeated,   since  it could  be  assumed
 that all engines  would  be at  the design targets  necessary  to
 meet  the  1987  standard.   Thus,  the  first step developed  the
 target-  levels  for  the  proposed  4.0/0.40  standards;  the  sane
 methodology  as  had   been  employed  for  the  6.0/0.60   proposed


standards was  used.   The  target  levels developed were  3.2-3.6
g/BHP-hr NOx and 0.30-0.33 g/BH'P-hr particulate.

     Next  the  analysis  assessed  the  effectiveness  of  various
control techniques.  Technologies  projected to be  required for
compliance  with 4.0/0.40  engine-out  emissions standards  were
broader  than   those   anticipated   for  the  6.0/0.60  proposed
standards  and  included  the  following;  additional  injection
timing   retard,   advanced   aftercooling   designs,   improved
combustion  chambers,  high pressure  fuel  injection,  exhaust gas
recirculation,  electronic  controls,  in-cylinder  heat retention
and  fuel  modification.   In  light  of  the  wide variation which
exists between  specific diesel engines,  it  was anticipated that
manufacturers  would,  on  an  engine  specific  basis,  select the
combinations  of these  technologies most  appropriate for  each

     Finally  in the  third  step,   the effect  of  the  proposed
standards  on fuel  economy was  examined.   In  the  short   term,
i.e.  immediately following  the  effective date of  the  proposed
standards,  the  projected  effects  of  the  4.0/0.40  proposed
standards  on fuel  economy was  for a  1-2  percent  penalty which
should be eliminated with time.

     3.    Trap-Based Particulate Standards

     In the  NPRM analysis[1]  of  the feasibility  of particulate
trap-oxidizers  for  heavy-duty diesel  use, EPA determined that
traps would  be  feasible for  1990 model year  HDDEs.   Due  to the
limited  amount  of  available  HO  trap  development data,  the
analysis   first examined  light-duty  trap   status  and  then
considered  the  degree  of  additional development effort required
by  the  heavy-duty   industry.   As  a  result   of   this  and the
ongoing  research  and  development  data,  EPA concluded  that
trap-oxidizers  would be  available  to  permit  compliance  with   a
HDDE  trap-based particulate  standard;  the  standard  level was
also calculated in  this analysis.   The following will synopsize
the  four steps  of   the  NPRM analysis:   LD  trap  status;   LD/HD
differences; HDD trap status; and emission  levels.

     Based  on  past  EPA analyses  and a contracted study [3,4,5],
the  Agency concluded that light-duty  trap technology was at   a
very   advanced  stage  of   development  and   light-duty  trap
oxidizers  would be  technically  feasible  no later  than the 1987
model  year.  The  findings that  traps were  feasible for 1987
model  year  LDVs  was  also  based  on Daimler-Benz's plans  to
certify  a   trap-equipped  vehicle  to meet  California's  1985
emission   standards.    Although1  there  were   still  unresolved
problems  associated  with  some trap  systems (e.g.,  introduction
of  a  fuel  additive  to  the fuel  to   induce  regeneration, the
development  of  a fully automated  positive  regeneration  system,
and   the   occurrence   of   increased   sulfate  emissions  from


 catalyzed  traps),   EPA  believed  that  the  other  manufacturers
 were   not   far  behind   Daimler-Benz's   trap  development   and
 compliance was possible for 1987 LDV use.

      The second step in  the  analysis  examined the applicability
 of  light-duty  trap  technology to heavy-duty engines, concluding
 that there was nothing  preventing the  adaptation  of light-duty
 technology,  with  additional  development,  to  heavy-duty usage.
 The  further  advanced  light-duty  trap  technology   formed   the
 basis   for   the   development   of  similar  technology   for   the
 heavy-duty   diesel   engine    industry.    However,   conditions
- specific  to  the HDDE environment  were identified which must  be
 considered  in  the  design of heavy-duty  trap,  oxidizers.    The
 major  light-duty/heavy-duty   differences  included:   engine  size
 and  load  factor;   operating  conditions  and   temperatures;   the
 useful  life  of  the engine; and  ash  accumulation.   Although
 considerable development  effort was found  to  be required of  the
 heavy-duty   industry in  the  adaptation  of   light-duty   trap
 technology   for  heavy-duty   use,   EPA  did  not  consider   the
 problems to  be without engineering solutions.

      The  analysis   continued with   a   survey   of   the  ongoing
 heavy-duty  trap  research  and development.   The Agency  found  a
 definite  lack of   data   from -the  HDDE  industry,   noting   the
 difference between  LD  trap progress,  where the  LD  industry  has
 had  to  work   towards    a  trap-based   standard,  and   HD   trap
 progress,  where the HD  industry  has not  had  that incentive.
• The  limited  development  work  was  primarily  focused  on  trap
 regeneration and  its control.  (Trap  type, for  the  most part  a
 direct  derivative  of  light-duty  design,  was  not  considered  a
 major   obstacle,  although some   design effort   in this   area
 remained).   Regeneration methods  being  evaluated included,  but
-were  not  limited  tc,  burners,  fuel  additives  and catalyzed
 traps.   Development of  an automatic  regeneration  control  system
 appeared  to  be the next major  step.   EPA realized that  traps
 were  not  at  the   time  a viable  particulate  control,  but  the
 Agency  firmly  believed  that,  with industry's  vigorous  pursuit
 of  a  trap-oxidizer  system,  traps  would be achievable for  HDDEs
" by  1990.

      The   final   step   identified  a   feasible  standard   for
 trap-equipped  heavy-duty diesel  engines.   The  proposed  trap
 standard   was   dependent  on   the    following   factors:     the
 engine-out  design  target  level,  the  deterioration  factor  (DF),
 the SEA adjustment  factor and the  trap efficiency.  The  target
 level,  SEA  adjustment  factor and  DF  for the  engine-out  emission
 level  of  0.60  g/BHP-hr  were  determined  in the  non-trap  standard
 section of  the analysis.  The  1.0  DF for  traps was  based  on 1DD
 particulate  emission tests of  over  50,000 miles  that  resulted
 in  no  significant  deterioration.   The  final  and most  variaoie
 factor-,  the  trap  efficiency,  ranged  from  50  to greater than 90
 percent,  dependent  on  trap  type.   The  Agency determined  that


with 80 percent efficient traps, HDDEs could  comply  with a 0.25
g/BHP-hr  standard;  with averaging, approximately  70 percent of
the HDDEs would require traps.  If  essentially  all vehicles are
equipped  with  90  percent efficient  traps  then a  0.10  g/BHP-hr
standard was determined to be feasible.

     The  technology feasibility  analysis  concluded  that  there
appeared  to  be  sufficient  time  for the manufacturers to design,
develop,  and prepare  trap-oxidizers for  1990 model  year HDDEs.
This followed  from  the fact  that  traps will be in production on
many light-duty diesels  no  later  than 1987.  The  five  years of
leadtime  between  mid-1984  and late  1989  were found  to  allow
adequate  time  for  the  additional  design  effort  required  -for
HDDE modifications.

     B.    Summary  and Analysis of Comments

     1.    Near-Term NOx and Particulate Standards

     The  proposed  6.0/0.60  standards  for 1987  represented an
attempt  by  EPA to  obtain  meaningful, yet  balanced,  reductions
in  both  NOx  and  particulate  in  the  near term.  For  example,
greater  NOx reduction could  have been   proposed.   California
already  has  a  5.1  g/BHP-hr  NOx  standard  for  HDDEs.  However,
California   has  no   particulate   standard   for   HDDEs   and
particulate  levels  almost  certainly average  well   above   0.60
g/BHP-hr.   Since   EPA  also  desired  to  establish  near-term
particulate  control,   NOx  controls  were  proposed only  to the
point  where they  did  not  unduly  impact potential near-term
particulate control levels.

     Though  always  intertwined,  the   issues  of feasibility and
leadtime  are more separate  here  than in  many  other cases, due
to  the fact that  the  6.0/0.60 standards  were  proposed to  take
effect  in a very  short period of  time,  just over  two years  from
the  date of proposal.  Thus,  those issues related primarily to
feasibility will  be discussed  first followed by those concerned
primarily with  leadtime.

     a.    Feasibility

    • Overall,  the 6.0/0.60  standards  were fairly well  received
by  manufacturers.  A  number  of   manufacturers indicated   that
they   were    feasible  for   either   1987  or   1988.     Most
manufacturers,  however,  took  issue  with  the  details  of EPA's
ana-lysis  in  support of the standards.  Thus, these  details  need
to  be  addressed,  as  well  as overall  comments  with  respect to
feasibility  and leadtime.

     These  details  fall  into three basic  categories.   The  first
 is  the  identification of  the design goal,  or target,  associated
 with  the two  standards.   The  second  deals  with the  projected
 effectiveness  of control  technology  and  the  ability to  reach
 the design  targets.  The  third deals  with  the effect of  these
 technologies on  BSFC, or  fuel economy.

     i.     Design Targets

     Design targets  are   a  function  of:     1)  the emission
 standard,  2)  the DF  applicable over  the  full  useful  life,  and
 3)  emission measurement  variability.   The model,  or  equation,
 used   to   determine   a   low-mileage   target   based   on   these
 parameters  is well  known  and accepted.  The only  issue relating
 to   the model   itself  is   the  assumption   that  the emission
-variability of   an  engine  is known sufficiently  well to  allow
 use of  the z-statistic  as  an  indication  of  the   statistical
 effect   of  this  variability  as  opposed  to   the  K-statistic.
 Therefore,  differences  in estimated design targets  arise due to
 the use of different  input  DFs  and  emission  variabilities  or
 the use of  the K-statistic  rather  than the z-statistic.

     A   substantial  amount   of  comment   was   provided  on  the
 development of  the target levels  necessary  for compliance  with
 the proposed  1987  standards.   Six  commenters provided numerical
 value   comments   on  the   low  mileage emission  target   levels
 necessary  for  compliance with  the proposed NOx  and particulate
 standards of 6.0 g/BHP-hr and  0.60 g/BHP-hr, respectively.   The
 target  levels provided by the commenters are  shown below:
                       Low mileage Target Level (g/BHP-hr)
 Draft RIA







              4 . 94-light heavy
              4.90-medium heavy
              4.84-heavy heavy
              4.5 to 4.9**

              4.88-light heavy
              4.84-medium heavy

              5.45 for
              product ion

0.42 (assumes improved

0.35-light heavy
0.29-medium heavy
0.21-heavy heavy

0 . 32-0 . 47-med.ium heavy'

0.32 to 0.37**

0.36-light heavy
0.30-medium heavy

0.47 (0.05% sulfur in
 * «
Depending on assumptions on DFs and variability.
Target levels nay be increased as more knowledge is gained.

     The  low  mileage  NOx  target  level  was  estimated  in  the
 Draft  RIA to be 5.1  to  5.5  g/BHP-hr,  based on a coefficient  of
 variation (COV)  for NOx and  particulate of 10 and  10-15  percent
 respectively,  and   full-life  DFs  of  0-0.48  g/BHP-hr  NOx  and
 0.04-0.06 g/BHP-hr  particulate  and  use  of   the   z-statistic.
 Neither  Cummins  nor Mack provided details on  the  methodology
 used   in  developing  their  target  level  estimates,   so  their
 estimates  cannot  be  technically  evaluated.   However,   both
 estimates are inside the range  identif-ied by  the Draft  RIA,  at
 least  for NOx, so  their estimates  of  the NOx DFs  and  COVs  must
 be  close  to  those of the Draft  RIA.

     The  EMA, GM  and IHC estimates  were  based  on  use  of  the
 K-statistic  to account  for  emission variability,  which  assumes
 that the  standard deviation  of  NOx emissions for a  given engine
 family is  unknown.  As discussed in  the Draft RIA,  the  more
 appropriate  statistic is the z-statistic,  since  fairly accurate
 estimates of  the  standard  deviation will  be available prior  to
 production  decisions  for  1988.   For  example,  as  evidenced  by
 Cummins'   comments,  manufacturers  are  already  testing  their
 production  audit engines  for  particulate.  No  new  information
 was received which justified changing  this conclusion.

     The  DFs  used by  EMA  and  IHC  were derived from  in-use
 engine data from two sources:   1) an  EEA study (performed  for
 EPA)  of  vehicle testing performed  at  SwRI and 2) engine  testing
 from   the.   joint   EPA-EMA   in-use  test   program.   A   linear
 regression   was   performed  on  the  after-maintenance  data  (or
 as-received  if  maintenance  was  not  performed)  from  these  two
 programs  vs.  mileage  to  derive  DFs   for  NOx and  particulate.
 The resulting NOx  DFs were  not  far from  those estimated  in  the
"Draft  RIA,  but the particulate  DFs were substantially larger.

     Generally,   such   regressions   are   performed  to   derive
 estimates of average in-use  emissions.  This was the purpose  of
 the EEA  study sponsored  by EPA.   Included in  the results  of
 such   a  study  is  an estimate  of how fleet-average  emissions
 change with  mileage  (i.e.,  an  in-use  average  DF).   However,
 unless the  engines  or  vehicles  tested  meet   the  criteria  for
 inclusion  in  a  recall  action,   the  resulting  DFs   are  not
 appropriate  for  use in a design target analysis.            -

     An  analysis of  the engines  included in these  two programs
 shows  their  condition  to  be far  from satisfactory for recall
 evaluation.   Many were tampered and  restorative  maintenance was
 performed on  only  13  of  48 engines.   Thus,  the  resulting  DFs
 essentially   represent   in-use   DFs    and   not  those   •.:"
 well-maintained  engines and  should not be used here.


     The GM DF  for  NOx  was estimated from  a  subset  of the data
referenced by EMA  and  IHC.  However,  even given  this fact,   it
fell into  the  range of  the Draft RIA.   GM's  DF for particulate
was simply estimated to  be 0.15 g/BHP-hr.   This is well outside
the range  used  in  the  Draft RIA, but  cannot  be evaluated since
its basis is not known.

     The Draft  RIA  NOx  DFs were  based on  1984 half-life data;
doubled  to represent full-life  DFs.   Full-life  1985  data   are
now  available   and  are  shown   in  Table   2-10.    Only   the
manufacturer-developed  DFs are shown,  since  they  are based  on
actual durability testing.  Assigned DFs  are  provided by EPA  at
the manufacturers choice,  but  these are worst-case estimates  to
encourage  actual  durability   testing.    Overall,  half  of   the
developed  DFs  are  zero and only three  are  significantly more
than the upper  estimate used  in the Draft  RIA (0.48  g/BHP-hr).
The average  developed  DF  in  each subclass is  0.0 (LHDDE) ,   0.1
(MHDDE)  and 0.32  g/BHP-hr  (HHDDE).   Thus,  the  range  of   the
Draft  RIA  appears  somewhat conservative for  LHDDEs and MHDDEs.
Since  quite  a  few  HHDDE DFs are  quite  near  0.48  g/BHP-hr,   the
Draft  RIA  upper  limit  appears   quite  appropriate   for  these
engines.   It should be  noted  that  manufacturers currently have
little  pressure  to  reduce  NOx  DFs   since  the  10.7 g/BHP-hr
standard  is well   above  low-mileage   emission levels.   Thus,
current   DFs,   particularly   the   largest,   could  very  well
represent conservative estimates  of  future DFs when they  become
a factor with respect to compliance.

     Lacking data,  the  Draft  RIA assumed  the DF for particulate
emissions  would be similar  to  the  NOx  or  HC  deterioration
factors  when  expressed   as  a  percentage  of  the emission  level.
Commenters contended that  normal  wear  in such  components  as  the
^11 o 1  -in^^^^Ton  r\i i mm  ^ *• e  r^rNnf-yole  in-i^^t*/^r*c  a n H n T c t- r% n T-Tnnc
^ «» >« *  .».*J >»*•*..***.»  £-«••••£-, *****  WV..WKW~W, «.».JWW»W — w  «»..*_ ^- _ w w .w . . _ _ . . -3 v
would  be expected to cause an  increase in HC and/or  particulate
emissions  while  causing a decrease  in  NOx  emissions.  Given  the
fact that  most  NOx  DFs  are zero  or  negative  and  this would  not
be expected  for particulate,  the use  of  HC  DFs  as a surrogate
is  probably more  appropriate.   Referring to  Table  2-11,   for
LHDDES,  the  ratio  of the  mean  deterioration  factor to the  mean
low mileage emission level was  found to be 0.14.  Corresponding
ratios  for MHDDEs  and  HHDDEs were  found to  be  0.05 and 0.06,
respectively.   Under  a   particulate  standard  of  0.60 g/BHP-hr,
the  low  mileage  level  will   be  roughly 0.5  g/BHP-hr  and  the
preceding  ratios developed from actual HC  deterioration  factors
would  correspond  to particulate  deterioration f-actors of 0.07,
0.025,  and 0.03  g/BHP-hr for  light,   medium  and  heavy  HDDEs,
respectively.   These values  bracket very  closely the DF  range
(.04 to  .06  g/8HP-hr)   developed  in the  Draft  RIA.    Thus,  tnis
range  continues to  appear  appropriate.

                              Table 2-LO

             1985 Federal Full-life Deterioration Factors

LHDDE                               DF (g/BHP-hr)CI]

 GM                                      0.0
 IHC                                     0.0


 GM                                      0.0,  0.0
 Caterpillar                             0.02
 IHC                                     0.0,  0.0,  0.61


 GM                                      0.65, 1.14
 Caterpillar              -               0.0,  0.0,  0.47
 Cummins                                 0.07, 0.39,  0.46,  0.46
 Mack                                    0.00, 0.00,  0.00,  0.37
 Volvo  White                             0.50
      Only manufacturer-developed  DFs are  shown.   Assigned  DFs
      are  essentially  worst-case  DFs  and are  not  necessarily
      indicative of an  engines actual DF.

                                 Table 2-11

                 1985 Model Year HDDE HC  Emission Levels and
              •Deterioration Factors  Developed by Manufacturers

 Manufacturer                   HC  DF            HC Low-Mileage Emissions

    General Motors               0.15                    0.53,0.46
    International Harvester      0.00                    0,79

 MHDDE         ^  '
    Caterpillar                  0.06                  •  0.62
    General Motors               0.05,0.00               0.58,0.84
    International Harvester      0.00,0.08,0.03          0.70,0.85,1.32
                                                       t              	.
•—- -Caterpillar	,  .	"  . . 0.00,0.21,0.01          0.19,0.36,0.32
    Cummins          -            0.00,0.00,0.00,0.02     "0.46,0.62,0.92,0.52
    General Motors               0.00,0.00               0.48,0.54
    Mac*                         0.19,0.00,0.00,0.00     0.90,0.69,0.74,0.54
    Volvo White                  0.10                    0.81,1.15


      With  respect  to  the  last  pertinent   factor,   emissions
 variability,  EMA,  IHC, and GM all used estimates which included
 lab-to-lab  variability.   This  would  be   appropriate   in  an
 analysis focused  on  pre-production certification  requirements,
 if  EPA   were  to  perform  confirmatory  tests  at   its  own  lab.
 However, the  focus  here  is  SEA, because  its requirements  are
 statistically more stringent than  those of certification.   SEAs
 are  performed  at manufacturers'  own  facilities.   Thus,   any
 differences between  a  manufacturers'  own labs  are  well  known
 and characterized.   Thus,  inclusion of lab-to-lab variability,
 particularly  insofar  as  these  estimates  were  based  on  the
 variability  among seven  independent test  facilities,  is .not
 appropriate  here.   When  this  is   taken   into   account,   the
 estimates of EMA,  IHC, and GM would be very similar to  those of
 the Draft RIA.

      Overall,   then,   the  inputs parameters  estimated  in  the
 Draft RIA  still appear appropriate.  Thus,  the  design  targets
.remain unchanged  at  5.1-5.5  g/BHP-hr  NOx  and 0.47-0.51 g/BHP-hr
 particulate.     However,   the  facb  that  most  manufacturers'
 estimated design  targets  were well  below  these levels should be
 kept   in  mind  below  as  control  technology  effectiveness  is
 discussed.  An  unrealistically  low design target  overestimates
 the  degree  of   control   necessary  to   achieve  a   standard.
 Therefore, either  the necessary application  of  technologies is
 overestimated,  or  a  standard is termed   infeasible  when it  is

      ii.   Control Technology Assessment

      The  analysis  of  HDDE  control  technology   is   inherently
 difficult,  because  each  manufacturers'  engines  are  designed
 somewhat  differently and  have  varying technical  capabilities.
 Differences between  the generic  HDDE subclasses compounds  this
 task.   Thus,  engine-specific analyses are  not possible  due to
 the complexity  of  the  task.   However, even  if such  an  attempt
 were   possible,  the  necessary data  are   not  available  in  most
 cases.  Thus, the analysis in  the  Draft  RIA  and that performed
 here   must  address  generic  control  techniques   and  reduction
 capabilities, while  at  the  same  time  considering  differences
 between engine designs  insofar as possible.

      Another  factor  adding to  the  complexity of  the  task is the
 rapid  change  in   technology currently  affecting  HDDEs.  • New
 technologies, such as  enhanced  aftercooling,  variable injection
 timing,    electronic   engine   controls  (EEC),   higher-pressure
 injection  and  higher efficiency,  faster  response  turbochargers
 are  all  being  introduced  to   some  degree  to    improve  BSFC,
 regardless   of    emission   levels.    However,  many   of  these
 technologies  also directly effect   NOx and  particulate  and are


 among    those   considered    below    as    potential    control
 technologies.   A problem is tha€ all  of  these can  be  optimized
 for  BSFC or emissions and  interact  in a complex way.   Thus,  it
 is  also difficult  to  determine a  pre-control  baseline.   The
 result  is  that future  technology  must  be estimated  both  with
 and  without these  standards and  data  from engines  encompassing
 a  representative sample of these  technologies  must  be  relied
 upon  to   estimate   overall   control   effectiveness.    While
 important   here,  these  factors  are even more  dominant  in  the
 analysis of the 4.0/0.40 standards  to  follow.

     Unfortunately,  little  data  were  received  in  comments  on
 the  1987 standards which quantified  the  effect of  the various
 control  techniques  projected to  be both available  and  effective
 in  achieving these  standards  in  the Draft RIA.   Most commenters
 simply   stated   whether   or  not   the   6.0/0.60  standards  were
 feasible    and,   if   so,  when.    Some  also   presented   their
 qualitative  judgment  of  EPA's  feasibility  analysis.   A  few
 (e.g.,  GM)  presented  charts of NOx/particulate   and  NOx/BSFC
 curves  for  each of  their engines.  However,  without  test  data
 and  descriptions,  these also  cannot  be  properly  evaluated.
 Manufacturers'  comments  pertaining  to overall  feasibility  will
 be  summarized  first,  followed  by  general comments,  pertaining
 to   the   Draft   RIA  analysis.   These  comments  will  then  be
 analyzed   using what   data  were supplied,   as  well   as  those
 included  in the Draft  RIA.   -  .              ...

     Daimler-Benz   stated,   without  qualification,   that  their
 MHDDEs  could achieve  compliance  with  the  proposed  standards in
*:1987,  as  did  Volvo  White  with  respect  to  1988.   Ford  also
".stated  that" compliance  with  the  proposed  1987 standards  was
• achievable,  but   indicated  this   conclusion   was   based   on
 projections  chat  both  DFs  and  emission  variability  would  be
 relatively  low  (which  they   expected   and   which   appeared
 reasonable  given the  analysis  presented  above).   GM  indicated
 that its  medium- and  heavy-HDDEs  could also comply  in  1987,  but
 with some  fuel economy penalty  (which is  addressed below).   in
 the  case  of  its   light heavy-duty  engine,   GM indicated  that
 compliance with both the proposed  particulate and NOx  standards
 was  not  achievable  simultaneously.

     Comments by the other  manufacturers  as  well as by EMA did
 not  include  direct  statements on either  an  anticipated ability
 to  comply nor an anticipated  inability to  comply.   However,  the
 comments  did  include  discussions  of   the -technologies required
 for  compliance  and  the  time required  for  implementation.   Thus,
 it   is  reasonable  to  infer  that compliance  with  the proposed
 standards  was  considered to be  technically achievable by these
 other  manufacturers,  as  well.   Cummins   and  Mack  did mentis
 implementation   years  of  1989 and  1990   respectively,   for  j-.
 least  'the  NOx  standard.  However,  leadtime will  be considered
 further  below.


     Overall, the only  manufacturer  to absolutely  question  the
feasibility  of  the  6.0/0.60  standards was  GM  for  its  LHDDEs.
On the surface,  this is rather surprising since  data  generated
in  EPA's Ann  Arbor  Lab on  a  low-NOx version  of  this  engine
(referenced  in  the  Draft  RIA)  showed it  to  have the  lowest
combination of NOx  and  particulate emissions of any engine (3.0
NOx  and  0.46  particulate,   g/BHP-hr).   Also,  prototype  data
submitted by GM  after the  original proposal (then confidential,
but  recently made  public   in an  EPA-sponsored  study[6]>  show
emissions  to  be  4.1  NOx/0.46   particulate  and  2.8  NOx/0.52
particulate  at  two  calibrations  (all  in  g/BHP-hr).  While  the
levels  of  1984   production  engines  are  somewhat  higher  (4.2
NOx/0.66    particulate     and    3.6    NOx/0.62    particulate,
respectively),  these levels  are  still  low  relative to  those of
the  other engines  and   no  incentive  existed  in  1984  to  keep
either NOx or particulate as  low  as the prototype levels.

     GM  did  not  refer  to  any of  these data,  but did  present a
NOx/particulate trade-off  curve  for  this  engine.   The  curve is
slightly  below   the  1984  production  data,   but  well  below
prototype  curve.    No   explanation   is  given  concerning  the
prototype/production  difference.   Also,  GM's  estimated  design
target  for  the   particulate   standard is  0.32-0.36  g/BHP-hr,
which  is below  even the  prototype  data  and  may  explain  GM's
conclusion.   It   should  not  be  necessary,  based  on   EPA
estimates,  to  design   an  engine  below  a  design  target  of
0.47-0.51    g/BHP-hr    particulate,    as   discussed    above.
Consequently,  this   engine  must   be  considered  capable  of
complying with the proposed standards.

     Moving on to  comments on the Draft  RIA analysis,  a  number
of  manufacturers  (Caterpillar,   in  particular)  indicated  that
some of  the  analyses were rather  simplistic and not realistic.
For  example,  Caterpillar  took  issue  with EPA's statement that
California's 5.1 g/BHP-hr NOx standard could easily be met with
simple  injection  timing  retard.   Caterpillar   also  disagreed
with EPA's implication  that  transient  particulate emissions can
be  reduced to  steady-state  levels,  through improved transient
fuel rate control.    •  •

     With  respect  to the first statement,  Caterpillar  took the
statement  more  literally  than  intended.   The  primary  point
being  made  was  that,  with respect  to  techniques  designed
primarily  to control  NOx control  techniques,  injection timing
was  sufficient  (i.e.,   no  other  NOx  control  techniques  were
required)  and  the   point  was  not  that  absolutely  no  other
changes  (e.g.,  recalibrations) would  be  required.  Caterpillar
lists  a  number  of   changes  made  to  its  California  engines  in
addition  to  injection   timing  retard,  such  as  power  de-rate,


 turbocharger  modifications,  and  fuel  governing  modifications.
 These  are  reasonable  recalibrations  whenever  a  basic  engine
 parameter,  such as injection  timing, is  changed.   However,  they
 in  themselves   are   not   necessarily   NOx   control  techniques,
 though  their  cost  must  be  considered.

     With  respect   to  the   second  statement,   Caterpillar's
 judgment  is  based on  a  belief  that  advances  in  turbocharger
 design  have  already  achieved most  of  what  is  to be  gained " in
 improved   transient  response.    Further,   they   claimed   that
 over-fueling  is necessary  to accelerate an  engine.   Again,  the
 point  being  made  in  the   Draft RIA  was  not  that  the  entire
 transient/steady-state  difference could be  eliminated,  but  that
 improvements      were      possible      and     the      current
 transient/steady-state  difference  was  an  indication  of  this
 potential.    Given the work  known  to  be  underway  by -both
 turbocharger   manufacturers   and  other   HDDE   manufacturers—
-evidenced by  the  numerous  technical  papers  in  the   area  even
 though  most  of  what  is being  done  is  proprietary—it does  not
 appear  reasonable  to  conclude that  turbocharger response cannot
 be  measurably  improved.   Also,  the  potential  capability  of
 electronics   to  precisely   limit  fuel  delivery  to minimize  any
 particulate   control/performance  could  also  be  substantial.
 Whether   such improvements  can-  be  achieved  by  the  1987-1988
 timeframe fleet-wide  is another issue.

     Each manufacturer  also   identified,  in  varying  degrees' of
 detail,   the  technologies   which  it  expected  to  use  on one  or
-more of  its  engines  to  achieve compliance  with the proposed
 6.0/0.60   standards.    The  technologies   identified   were   as
 follows:    1)  application  of  turbocharging,  2)  turbocharger
 modifications,   such   as   improved   efficiency  and  transient
 response, 3)  addition  o£   a£tercooling  to  turbcchargsd engines,
-4)  enhanced  aftercooling),   5)  injection  timing  retard,   6)
 addition   of  variable  injection   timing,   7)   increased  fuel
 injection  pressure,  8)   fuel  injector  modifications,  and  9)
 modifications  to  the  combustion chamber   and  air  swirl  rate.
 Manufacturers also indicated that  they  anticipated  an ongoing
 introduction  of  electronic  controls   focused  mainly  on  the
 minimization of fuel  economy penalties.

     These  technologies   are   basically   the  same   as  those
 projected  in  the   Draft  RIA  for  both  the  1987  and  1990
 standards.   While some use of the technologies associated  witn
 the  latter  standard  was  anticipated  in. 1987,  manufacturers
 appear  to  be  utilizing   a  greater  number  of  combinations  of
 technologies at the  6.0/0.60  level than  had been  projected in
 the Draft RIA, possibly because of fuel economy concerns.


     Telephone   communications   with   manufacturers  concerning
their   1985   model   year   California   engines   showed   that
combinations  of the  technologies listed  above are  in use  on
these  engines.   However,  since the  half-life California  NOx
standard  is  essentially   equivalent  to  a  5.1-5.35  g/BHP-hr
full-life standard,  it is 0.65-0.9 g/BHP-hr more stringent than
the  proposed Federal  standard  and  not all  of  these technology
modification/additives  (at   least  these  which  are  NOx related)
should be required  to  comply with the  6.0/0.60  standards.

     With  respect  to  the  NOx  standard,  EPA  acknowledges that
sole  reliance on injection timing  retard to achieve compliance
in  the  6.0  g/BHP-hr  NOx standard  could result  in significant
fuel penalties.   Thus, to minimize fuel penalties manufacturers
may  elect  to  increase  the  use of  aftercooling  and  variable
injection  timing.   However,  the 'use  of  enhanced aftercooling,
particularly air-to-air  units,  appears more  appropriate  at NOx
levels  more  stringent  than  6.0 g/BHP-hr,  it  should not  be
necessary at 6.0 g/BHP-hr NOx.   If air-to-air  aftercooling were
applied,  it would be  to  reduce  BSFC  and should not be included
as a cost of the 6.0 g/BHP-hr standard.

     With respect  to  particulate emissions,  some additional use
of  turbocharging was  projected in the  Draft RIA,  particularly
with respect to  Caterpillar's  3208  engines.   This was confirmed
by   Caterpillar  in   their  comments.   Also,   the  Draft  RIA
identified   the  general  need   for  modifications  to  existing
engine  components,  but  none   involving  additional  hardware.
These  components   include   modified   injectors and  combustion
chambers,   improved   fuel   governing   during  transients,  and
moderate  increases  in  injection  pressure,   all of  which  are
described in more detail  in the  Draft  RIA.

     Due   to   the  difficulties   mentioned   above,   such  as
heterogeneous engine  designs,  lack of engine-specific  data and
rapidly  changing technology  to  reduce BSFC, specific estimates
of  the  technological  changes  necessary  for  each  engine  cannot
be   made.     However,  most  of  the  changes  described  above
primarily   involve   research,   development  and  tooling.   The
revised  components should   inherently be  no more  expensive in
the  long-run than  the original components.   Thus,  the cost of
these  standards  may  not   depend  strongly  on  the  number  of
changes made, but  rather on the need to  perform the necessary
research  and development  to  determine which  changes  actually
.need  to  be made.  For the  most  part,  much of  this research has
been ongoing already or  performed.


     iii.   HC and Fuel Economy Effects
     No   technically   supportable   comments   were   received
indicating  the  1988  standards  would significantly  increase HC
emissions.   However,  comments  pertaining  to  the fuel  economy
effects  of  the   proposed   standards  were  provided  by  most
commenters.  The estimated  fuel economy penalty  anticipated by
each comrnenter  are shown below together with  the basis  for the
estimate,  when one was provided.

	Commenter	     	Fuel Economy Penalty	

Caterpillar                  3-12    percent   -(1985    Federal/
                             California difference)

Cummins                      1-3.5 percent    -           	

GM  ..  _   •                  3-5 percent for MHDDE
                             4-6 percent for HHDDE        '"'	
                             2 percent for new design HHDDE

International Harvester      4-8 percent from NAS study
                             5.4-7.2  percent MHDDE
                             7.7-8.3    percent    HHDE     (1985
                             Federal/California difference)

Mack                         6 percent (1985 Federal/ California
                             difference is 4.7-12.5 percent)

Daimler-Benz  •     ...     No significant loss  in fuel economy

v  '   Many   of    the  manufacturers'   projections   on   reduced
efficiency  of   fuel  utilization   were   based  on  differences
between  1985 model  year  Federal and California  engines.   This
is  not  an appropriate comparison since  the  California standard
is  5.1,  not  6.0 g/BHP-hr  and  the California  engines  are  1985
models,  not  1988.   The California  standard  is  a  full-life
standard,  but  does not  include an assembly  line test program.
This  allows a  somewhat  smaller  safety   margin,  since SEA is
generally  considered to  be  statistically more  stringent   than
certification.  However, this difference  should be only a  small
part of the 0.5 g/BHP-hr margin attributable to  SEA.  Thus, the
California  program can be  considered to  be  essentially  on par
with  the  Federal program at equal  standard levels.   As the 5.1
g/BHP-hr  NOx standard  is  much  closer  to - 5.0 rather  than 6.0
g/BHP-hr,  the  California data  are  more  useful in assessing the
fuel economy effect of the 5.0  g/BHP-hr standard  than this

     IHC  made an additional comment  that  the  NAS study  was not
based  on  "old"- technology,  as  concluded  in  the  Draft RIA, but
on  advanced engine  technology,  at  least  insofar as the data IHC
supplied.   Whether  or not  this is  true for other manufacturers'


data  cannot   be   determined,   since  no  other  comments  were
received on this point.  However, even if  some  of  the data were
from  advanced  technology  engines,  the   absolute  NOx  and BSFC
levels  of  the  NAS curve  and  the  resulting  trade-off  make  it
apparent   that  any advanced technology  was  applied to optimize
BSFC  and  power and  not  NOx  control.   This  approach   is  not
consistent with the  approach  taken in  estimating  the economic
impact of this  rule (see Chapter 3), where the  cost of advanced
technology  is  being  charged  co  NOx control.  Given  this,  any
fuel penalty  should  be determined  from  the BSFC of  the engine
without  the advanced  technology,  not with it.  Also, the nature
of the  NAS  study  made it  impossible to  cite the  specific data
used to  derive their  NOx/BSFC curves.   Therefore,  it cannot be
determined how much optimization  of BSFC  occurred at  low  NOx
levels.    Thus,  the  NAS  BSFC/NOx  curve  still   appears  to
overestimate  the effect of NOx control.

     There  is  another  reason why  some  of  the  manufacturers'
estimates  shown above  may  overestimate the fuel economy  penalty
of the  6.0  g/BHP-hr standard.   That is  the fact that  at least
GM  and  IHC  used   a  low-mileage  NOx design  target  0.6 g/BHP-hr
lower  than that necessary.   Use of  a  5.1-5.5  g/BHP-hr target
would  result   in  lower estimated fuel penalties.   This  appears
to be  confirmed by Cummins estimate.  Cummins' estimated a NOx
design   target  with  the  above   range   and  also projected  the
lowest   fuel   economy  penalty  of  any  manufacturer,   except

     While  1985  California  BSFC  penalties cannot  be directly
applied  to   the   6.0  g/BHP-hr   standard,  they  can  be  used
indirectly  to  confirm the  0-2  percent  estimate  of  the NPRM.
These  differences  between  the California  and Federal situation
need  to  be   considered.    One,   an additional  three  years  of
leadtime will  be  available  allowing  additional control system
optimization   .    Two,  the  low-mileage   targets  will   be  0.9
g/BHP-hr NOx  higher   so   less  NOx  control  will   be  required,
lowering any  fuel penalty.   Three,  the  technologies being used
in California are primarily quick  fixes,  requiring low  initial
capital  investment (research,  development, soiling).  Given the
longer  leadtime  available  and  the  fact  that  nationwide sales
will  be effected  by  high  BSFC,  and not just California sales,
much more  comprehensive  research  and  development,  resulting in
optimized  control approaches  and  lower  BSFC penalties,  should
result  even  with  today's  technology.    Thus,  the   upper  end of
the  current  California  penalties   must  be  considered  extreme
under  these  conditions.    The  lower end  of  the penalties,  3-4
percent,   should   also  be   able   to   be  lowered,  given  the
additional  leadtime  and  added  return  for the  same  investment
(i.e.,  Federal vs. California  sales).   Thus on average in the
short  term,  a maximum penalty of  2  percent  may result from the
6.0/0.60 standards.   The  possibility  of no penalty also exists
given  Daimler-Benz'  comment.    Thus,  in  the  short  term,  the
average  fuel  economy  penalty should  be 0-2 percent.


     In the long run, beyond  1988,  one would expect the penalty
to  disappear   altogether.    This   is   because   the   advanced
technology projected  in  the 1991 timeframe  should improve fuel
economy such that any short-term  penalty will be  eliminated  by
the  early  1990's.   Also,  general  improvements  in BSFC  will
lower fuel consumption over the emissions test  cycle and,  other
things being equal,  NOx  and particulate will decrease as well.

     It was estimated in  the  Draft  RIA that BSFC would decrease
roughly  1.5  percent  per  year  in  this  timeframe  based  on
comments  from  MVMA  and  EMA to  the  MOBILES  development process
confirm  this  figure.    Thus,   three  additional  years  should
provide a  4-5  percent  reduction in NOx  emissions  simply due to
BSFC improvements.

     b.    Leadtime

     The  issue  of  the amount  of  necessary  leadtime associated
with  the  6.0/0.60  standards  received  a  substantial  comment.
This  analysis  will  begin by  describing the  steps  necessary in
developing  engines   and  vehicles  to  meet   emission standards,
along with the time  associated with each  step.   The comments to
the  1987  implementation  date  will  then be  summarized,  followed
by an analysis of those comments.

     All work necessary for emissions  compliance by the engines
does  not  have  to  be completed prior  to initiation of design
work  for  engine integration into  the vehicles,  but sufficient
progress  has to be  made  in engine  development  so  as to clearly
define the engine envelope  (overall  spatial requirements of  the
engine, including aftercooling).  The primary tasks involved in
the  successful  development and marketing of  engines   complying
with  the  6.0/0.60  g/BHP-hr standards  in vehicles  are  shown in
Table 2-12.

     The  total  leadtime  requirement  for engine  development  is
the  sum of  tasks  A  through  H less  task  C, or  31-38 months.
Since the standards  are applicable  to  all HDDEs, the greater of
the  two  time  requirements  for durability  data  development  was
used.   In  the  case  of   vehicle   development,  the   leadtme
required  is  the sum of tasks  A,  B, C, I, J,  and  K, or a total
of 28-36 months.

     Starting with a March 15,  1985  date  for publication of  the
final  rule,  the  time  available  for   implementation of   new
standards  by the 1987  model   year. (January  1,  '1987)  would be
approximately  21 months.   The  time  available for  implerrentation
of new standards by the 1988 model  year (January 1, 1988) would
be  approximately  33  months.   Since  the   time  available   for
implementation  by  the  1987 model  year  is  significantly less



                           Table 2-12

            Leadtime Projection - 6.0/0.6 Standards


     Identify, by engine, technologies probably
     required for compliance, develop initial
Procure initial design hardware, build and
test initial design test engines

Develop engine envelope requirements for
vehicle builders

Develop second level engine designs

Procure hardware, build engines with "
alternative calibrations and develop
emission data and fuel economy
characteristics by calibration to define
durability engine calibrations

Develop emission durability data
Develop data from emission data engine

Coordinate emissions certification
compliance with EPA

Develop overall vehicle design considering
the effects of all engines offered in each

Procure new dies for the manufacturer of
redesigned vehicle components

Confirm mechanical durability of redesigned
vehicle components
Time Required

 3-4 months

 4-6 months

 1-2 months

 2-3 months

 8-9 months
 4-5   months
 for   light-
 months   for

 1 month

 2-3 months

 8-10 months
                                                     7-9 months*

                                                     5 months**'
Reference  2.
60,000 miles  at  40 mph average speed, two effective 7-hour
shifts per day an-1  six days per weeX.


 than  even the  minimum time estimated  above,  implementation  in
 the  1987  model  year does not.ap'pear  feasible  for most  engines.
 The  time  available prior  to  the 1988  model year is within  the
 range of  estimated  time  required.   Also,  the entire process  can
 be  accelerated   in  those extreme  cases   where more  time  is
 necessary.   Thus,   1987  should be  ruled  out  on the  basis  of
 inadequate  leadtime for industry-wide  compliance; however,  the
 1988 model year appears  feasible.

     Moving   to   the   comments,   one  commenter,  Daimler-Benz,
 stated  that  their  engines  could be brought  into  compliance with
 the proposed  standards  for  the  1987 model year.    GM stated that
 all  but their one  LHDDE could comply  in 1987.    However,  other
.commenters   indicated   that  a  greater  amount   of  time   was
 required,    usually   one   year,    but     occasionally    more.
 Specifically,  Ford and  Volvo  White acknowledged  that  1988  was
 feasible  while  Caterpillar indicated that 1988 was  the  earliest
 date   feasible.    International   Harvester  estimated   that   39
 months  starting from  the  date  that the engine configuration is
 finalized would   be   necessary  to  allow   integration   of  the
 reconfigured  engine envelope  into the  vehicle,   to  accommodate
 changes  in   engine  cooling  requirements,  the  addition   of
 air-to-air   aftercooling,   the  addition  of   electronics   and
 compliance with noise  and  safety standards.

     Cummins  and  Mack  requested  that the standards  be  delayed
 until   1989  and  1990,   respectively.   Both  cited the  statutory
 mandate of  4  years leadtime,  but also  referred to  technical
 difficulties.   Cummins  indicated  generally that anything  less
 than  the  statutory leadtime would require them to  accelerate
 development  of their  planned engine modifications  to  a  degree
 which  would  seriously  affect  the  durability,   reliability  and
 fuel  efficiency  o£ their  engines.   Mack  considered  the  1990
 date  necessary because  essentially all  Mack  engines had  to be
 redeveloped   and   personnel   limitations   precluded   earlier

     The  two projections  (by  Daimler-Benz  and   GM)  of   1987  as
 the  feasible  year  of  introduction  indicates   the  ability  to
 compress  the schedule  described  in  Table   2-11  above.    It  may
 also  indicate  that manufacturers  are  starting   from  different
 points  (i.e.,  levels of current  emissions).

     Without  emission  data or  specific  leadtime estimates,  it
 is  impossible  to  evaluate the  IHC,  Cummins  and Mack  leadtime
 estimates,  which   are   the  only  ones  requesting  time  beyond
 1988.   (Cummins'   and  Mack's  legal  agreements are  addressed in
 the  Preamble  to  the  FRM.)  Generally  speaking, the   types  of
 changes  being  referred to by IHC  should   not  be necessary co
 comply  with the 6.0/0.60  standards.  They may be desirable  in
 the  long-run  to  mprove  BSFC,  but they are  not driven  by the
 6.0/0.60  standards.


     Given  the  above  leadtime  analysis,  the  infeasibility  of
1987  as  an effective, date,  the general  support  for  1988  as  a
feasible  year  of   implementation,   and  no   clear,  supported
arguments  by  any manufacturers  against  it,  1988  is determined
to be the year the 6.0/.60 standards should be implemented.

     2.     Mid-Term NOx and Non-Trap Particulate Standards

     A  mid-term  (1990)  NOx  standard   of  4.0   g/BHP-hr  was
proposed  for  all HDDEs.   A  1990  particulate  standard  of 0.40
g/BHP-hr  was  also   tentatively   identified   as   the  non-trap
technological  limit,  and  proposed as a possible  standard for
non-urban  (linB-haul)  HDDEs.   In identifying these  levels, the
same  approach  was  used  as   that described  above  concerning
development of the near-term standards.   The  goal  was to obtain
both  NOx and  particulate emissions,   but NOx  emission  control
was  balanced so  as  not  to   severely  impact  the   ability  to
control particulate emissions.

     The  difficulties  in  performing  an  analysis  such as  this,
which were described with  respect  to   the  6.0/0.60  g/BHP-hr
standards  above,  apply even more here.  While the heterogeneity
of engine designs is  the  same, technology is  changing even more
dramatically  in  this  later  -timeframe  and  the   interaction
between  control  techniques is even stronger.  Also,  even less
data  exist than  was  available  for   1987  technology,  to base
feasibility judgments on.

     Again,  as  with  the 1987  standards,   there  are  two  basic
issues:   technical feasibility and  leadtime.   However, here the
issues  associated  with  leadtime  are  much  less   significant,
because  the  implementation  date  is sufficiently  distant  to
allow  significant  research  and  development  application.   The
long  leadtime   available  should  provide  manufacturers  with
adequate  opportunity  to  overcome   problems  and   undesirable
effects  associated  with  additional   NOx  control.    Also,  the
proposed  1987  standards  require  delay  until  1988  and  the Act
requires   a  three-year  interval  for NOx emission  standards;
thus,  the  mid-term  NOx  standards  cannot   take  effect   until
1991.   This analysis  will presume  simultaneous   implementation
of both  NOx and particulate standards since  that  will maximize
the  manufacturers'  ability to  design engines that can meet both

     a.    Technical  Feasibility

     The  analysis of these technical  comments  will  follow that
for  the  6.0/0.60  g/BHP-hr standards.   The  one exception  is that
no reanalysis of  design  targets will  be  performed here.    No new
information  is applicable  that  was  not  already  discussed with


respect to the 6.0/0.60 standards.   That  analysis  confirmed the
Draft RIA's targets.  Thus,  the  Draft  RIA NOx target of 3.2-3.6
g/BHP-hr will  be used below.   A target associated  with  the 0.40
g/BHP-hr particulate standard was not  explicitly  determined in
the Draft RIA,  but  it would  be  about  0.30-0.33 g/BHP-hr.   The
only point  to  keep  in mind  is  that  the  targets  manufacturers
used  in assessing  the  feasibility  of  the  4.0/0.40  g/BHP-hr
standards  are,   for  the  most  part,   significantly  lower  than
those deemed  necessary  here.  Therefore,  their statements  may
exaggerate the feasibility efforts  of  various standard levels.

     Every manufacturer  stated that the  1990  4.0 g/BHP-hr NOx
and    0.40    g/BHP-hr    particulate    standards    were   'not
technologically  achievable  with  any  combination  of  known  or
anticipated technologies.  While the analysis  of  the Draft RIA
for  both the  4.0 g/BHP-hr  NOx and  0.40 g/BHP-hr  particulate
standard identified a  number of potential control technologies
in  each case  and  used available data  to roughly estimate the
potential control  efficiency of each  technique,  no  commenters
presented parametric studies  of  any of these technologies which
would  better  demonstrate  their  potential effectiveness.   Many
general  comments  argued  against  the  efficiencies estimated in
the  Draft  RIA,  based  on technical  grounds,  but without  the
depth  of analysis necessary to fully  support  the  point being
made and negate  the  point  of  the Draft  RIA.  Thus,  while  a
degree  of  doubt has been  thrown  on the  estimates of  the Draft
RIA, insufficient  data  are  available to  go  through   the  NPRM
analysis  point   by  point  and  reestimate  the  effect  of  each

     However,    some   confidential   data   were   made  available
indicating the  combined effect of a number  of these techniques
(e.g.,   increased  injection pressure and  enhanced affcercooling).
as  well  as   the  projections  of  the  levels  feasible  in  this
timeframe.   while  such   data   cannot   be  used   to   directly
determine  the fullest  potential  of one  or  more  technologies,
they  do  represent  the  most  quantitative  set  of  estimates
available.  Some  degree of  evaluation can be  applied  using the
estimates  of  the  Draft  RIA.  To  facilitate  their use here,
these data have  been  combined into  a  single figure (Figure 2-2)
which   shows  both   current   levels   of    NOx   and  particulate
emissions  and the manufacturers anticipated  achievable  levels.
Superimposed   on   the   best   achievable   emission    control
projections are  the  range of  low-mileage targets  as previously
developed  for the  6.0/0.60  standards (points At   and  A,)  and
the  midpoint   of  the   low-mileage  targets  for the  engine-out
standards of 4.0/0.40 (point  E) .  A low-mileage target  level ::"
4.2-4.6  g/BHP-hr  for a  NOx  standard  of   5.0  g/BHP-hr   (vert.-31
lines  B) was  also  developed by  the same   procedure  and  i!-e
midpoint  shown  for  comparison  purposes.   Particulate  leveis L

             0.9 .
             0.8 .
             0.7 -
             0.5 J
             0.4 .
             0.3 -
             q.2 -
             0.1 -
                                                         figure 2-2                    ,

                              Ileavy-Puty Engine NQx and  Particulate Enqina-Oub Characteristics
CCurrent Engines

  Best Teclinologies
                                 2       J
7      8
        10      11
                                                         NQx Qiiissions g/DIIP-hr


and D  correspond  to the  intercepts  of lines  C  with  the  upper
and lower bounds of the  projected range of emissions using best
available   technology    combinations    as   estimated   by   the

     The  first  observation  to  make  about  the  best  technology
estimates is  that  the  upper  limit  barely  passes  through  the
design targets for  the  6.0/0.60  standards.   Thus, it represents
fairly near-term  "best  technology."   Second, the  lower  limit
does not  even  approach the  targets for  the  4.0/0.40  'standards.
Given this,  it is reasonable to evaluate the  lower  limit  of the
best technology curve  against  the projections made in the Draft

     First,  with respect  to  NOx,  the  Draft RIA  analysis  of 4.0
g/BHP-hr  NOx  relied upon large NOx reductions at constant BSFC
for   separate   circuit   and  air-to-air  aftercooling.    This
estimate  was  based  on  data  from  one  GM/DDA  engine  and involved
estimating  NOx  reductions  beyond that  evidenced  by the  data
based  on an  estimated   BSFC/NOx tradeoff  for   timing  retard.
Thus,  no  actual NOx  data  below  5.0  g/BHP-hr  were  available.
Complicating matters was  the absence  of any particulate data in
the study.  While it can be assumed  that  these  were  not above
0.80  g/BHP-hr  since BSFC was  improving,  particulate may have
been well above 0.60  g/BHP-hr.   Thus,  these data  may  not  be
inconsistent with the  curve in Figure 2-2, the problem may be a
lack of particulate control, not NOx control.

    • Based  on  testing   performed  on  the  same DDA engine,  the
Draft  RIA estimated that electronic  engine  controls  (EEC) also
had  the  potential  for   large NOx  reductions   at  constant BSFC.
Again,  however,  the  estimate   involved  assuming  a  BSFC/NOx
tradeoff curve for the  engine and  using  timing rstsrd s
BSFC improvemenc to estimate NOx  emissions at constant BSFC.  A
6.0 g/BHP-hr NOx level  was the lowest  actual  data point  in this

     Many manufacturers  stated   that  this  analysis  overstates
the  benefit  of  EEC.   Some  argued that  the  NOx   benefit  of EEC
depends on  the  final  NOx level   (i.e.,  its  benefit  is large at
6.0  g/BHP-hr NOx  and  negligible at  4.0  g/BHP-hr).   Others
argued that  the benefits of enhanced aftercooling  and EEC were
mutually  exclusive, due  to  the fact  that  combustion efficiency
limits the  use  of  either technology and  determining  when both
BSFC and  particulate emissions begin  to  increase dramatically.
Without   data  or   sophisticated   combustion  analysis,    it  is
impossible to prove or  disprove  these comments.   However,  these
data  could  be  consistent with that  in Figure  2-2,  given that
particulate emissions are unknown.


     With   respect   to   particulate  control,   the   Draft  RIA
analysis was  much  more  general and based on less data than that
for  NOx,  as  0.40  g/BHP-hr  was   not   the  primary,  proposed
standard.   The most feasible  technique  for  the  1990 timefrarre
was high-pressure fuel  infection.   The only data available were
steady-state  emissions  on  one engine.  The other technologies
were  1)  injection   rate  modification and/or  modulation  which
would  require significant  advances in  injector  technology,  2)
ceramics,  which  are   not  progressing  as  fast  as  some  had
projected  a year  ago,  and 3) conversion to methanol fuel, which
though  technically  feasible  requires  the  establishment  of  a
fuel  distribution   system   (except  for  buses,  as  discussed
below).   Otherwise,  small improvements  could .be  expected from
general  BSFC   improvements,  and  for  additional  optimization  of
injectors  and combustion chambers.   Thus,  little  data  exist
with  which to  refute   the  data  in  Figure  2-2  and  it  must  be
taken as  the  best  estimate of technology  in the  1991 timeframe
at this time.

     Given  this,  a  5.0  g/BHP-hr  NOx standard for  1991 appears
most  reasonable  instead of  the proposed 4.0 g/BHP-hr standard.
This  is  principally because of the adverse  tradeoff between NOx
and particulate which appears  likely.  While a 4.5 g/BHP-hr NOx,
level  may  be  potentially  achievable,  particulate  emissions
appear  to begin increasing  at a  distinctly higher  rate below
5.0  g/BHP-hr  NOx.   Also,   below   5  g/BHP-hr NOx,  particulate
emissions could in  some  cases  increase well above 0.6 g/BHP-hr,
making   trap   application   very   difficult.    This   would  be
particularly   true   with  respect   to   the  1991  0.1  g/BHP-hr
particulate  standard for  buses.   The BSFC  tradeoff  would also
begin increasing dramatically here, as well.

     With  respect   to  particulate,  the  range  of expected  low
mileage   engine-out  particulate   levels  corresponding  to  the
target  level  required  for a NOx standard  of 5.0  g/BHP-hr would
be 0.42-0.54  g/BHP-hr  (levels C and D in Figure  2-2).   If it is
assumed  that  trap-based  standards  are  implemented in  1991  and
1994  (i.e.,  pressure  to control  particulate continues through
the  1994  timeframe),   then  it 'is  likely  that   progress  will
continue  to  be  made in  reducing  engine-out particulate  levels
down  to  the  lower  best  technology' curve,  resulting   in  an
engine-out  particulate  level  of  0.43  g/BHP-hr  by  1994.   If
stringent  1991 non-trap  particulate standard  were   implemented
in  1991,  it should  also  be  able  to reduce engine-out levels to
0.42  g/BHP-hr,  but three years earlier.  Thus,  at best,  a 1991
non-trap  standard  0.50  g/BHP-hr appears  achievable.

      These  levels   are  at   least  partially  proposed  by  two
manufacturers.   Cummins recommended  (at  the  public  hearing  on
the  NPRM)  1992  target  standards  of  4.5 g/BHP-hr NOx  and 0.50


g/BHP-hr particulate.  Daimler-Benz  recommended  1990  standards
of  5.1  g/BHP-hr  NOx  and  0.68  g/BHP-hr  particulate.   Other
conmenters  either  did  not  recommend  any alternative  standards
to  4.0/0.40  g/BHP-hr  or   recommended   retaining  the  6.0/0.60
g/BHP-hr standards indefinitely.

     b.    Effect on Fuel Economy

     As  all  commenters   stated  that   the   4.0   g/BHP-hr  NOx
standard was  infeasible,  no  estimates  of   its  effect on fuel
economy were  made.   Nor  were any comments   received  addressing
the effect of a  5.0  g/BHP-hr  NOx standard.   However,  the  latter
can be estimated  from  the 1985  California/Federal  comparison
conducted above.

     As  in  that  section,   there  "are  a  number  of  differences
between the 1985  California 5.1  g/BHP-hr standard and  the 1991
Federal  standard  of   5.0  g/BHP-hr.    First,   the   Federal
low-mileage target  is  slightly   more  than  0.1- g/BHP-hr  lower
than that  in  California.  This would tend  to slightly increase
the fuel economy  penalty.   Second,  six  years of  leadtime  exist
between  the  California  and  Federal   situations  to  develop
improved technology.   Third,   any  adverse   BSFC  effects  would
affect  Federal  sales,   which  is  roughly 10  times larger than
California's.   Both  the available  leadtime  and   the  potential
national sales  impact,   manufacturers  would  be  expected  to  do
all that is possible  to  eliminate  any  BSFC  effect,  as opposed
to the California  approach, which is more short-term, quick-fix

     Given  that  significant NOx  control  technologies  such  as
electronics,  separate-circuit   aftercooling,  and   air-to-air
aftercooling  are  not currently  present  at  all  in  California,
and  given  their  projected widespread   use  by  1991,   it  would
appear  that  the   additional  leadtime   and  potential  national
impact would  overwhelm the  first.  Overall,  it would not  appear
unreasonable to project  the same  long-term  fuel  economy   impact
here  as  that  projected  in   the  NPRM  for  the  4.0  g/BHP-hr
standard, zero  percent.   However,  due  to  uncertainty  in this
analysis, and the projection  that the  BSFC/NOx curve begins to
turn sharply upward at  approximately a  5.0  g/BHP-hr  standard, a
long-term 0.5 percent  fuel economy  penalty may occur.   In the
short-term,   a  slightly  higher   1.0  percent  penalty  may  be

    " 3.    Particulate  Traps

     In  response  to  the  NPRM,  the  Agency  received  a  large
number  of  comments  directed   towards  its   heavy-duty   trap
feasibility analysis.   As  explained  above  in  the synopsis  of


the  NPRM  analysis,   the  Agency   concluded   that   traps  were
feasible  for  1990  model  year  HDDEs,  extrapolated  from  the
status  of  light-duty  trap  technology  and  the  design  effort
necessary  to adapt  this  technology  to  heavy-duty  usage.   The
comments  were both  supportive  and  critical  of  EPA's  analyses
and  conclusions.   This part  of the  regulatory  impact  analysis
will  respond to the  comments,  concentrating  on  the points  of
the   analysis  with   which   the   commenters   disagreed.    New
information   that   is  pertinent   to  the   heavy-duty   trap
feasibility  question  will  also be  incorporated.   The  comments
will  be  addressed  in the same  format  used   in  the  previous
analysis:     light-duty    trap    status;   LD/HD   differences;
heavy-duty  trap  status;  and  emission  levels.  .In addition,  the
leadtime issue will  also be addressed.

     a.    Light-Duty  Trap Status

     The  status  of  light-duty  trap oxidizers  was generally not
addressed by the commenters.  The  notable  exception was General
Motors.    GM's   position  is   that  technology  is  still  not
available  to meet the promulgated  1987 light-duty  vehicle  and
light-duty   truck  standards.    The  extensive  LOT  testing  (200
alternative  fuels and fuel  additives  combined  with over  150
trap  materials  in  over  500  traps)  conducted   by  GM  has  not
resulted  in an   identification  of   a  LD  trap  that  can  be
committed  to a production  program.  Thus,  GM strongly objected
to  EPA's  conclusion  that   light-duty  traps  are  technically
feasible for 1987  model year  vehicles.

     General Motors'  comments  notwithstanding/   the  Agency's
position   in  the   NPRM   was  borne  out   by   Mercedes-Benz's
certification of  its  3L  turbodiesel,  equipped  with a  trap to
meet  the  California  Air  Resources  Board (CARS)  1985 model year
standards.[8]   CARB's 1985  standard  for  California light-duty
diesel  vehicles  is 0.40  grams per  mile  (g/mi)  particulates,  to
be  further  reduced to 0.20 g/mi in 1986 and  0.08 g/mi  in 1989.
In  addition to  its  1985  California LDDVs (which are  also sold
in  other  western  states),  Mercedes-Benz plans to  add  traps to
all  its  U.S. sold  3L  LDDVS  in   1986,  a  year  prior   to  the
promulgated  1987  0.20  g/mi standard.

     Mercedes-Benz is not  alone in  certifying  a  trap-equipped
LDDV.   Volkswagenwerk AG  (VW)  plans  to install  a  trap  on its
larger  diesel  LDV  (Quantums)  in  California  beginning  in tne
1986   model  year.[7]    VW   intends   to  equip  - all  federally
certified  Quantums  with   trap-oxidizers the  following  year  to
comply  with the  1987  LDV  particulate standards.   The  trap
applications of  Mercedes  and VW are proof  that trap-oxidize..,
are  a  viable form  of light-duty  particulate emissions control.


      b.    Light-Duty/Heavy-Duty.. Differences

      The   manufacturers   which   commented   agreed  with   EPA's
 analysis   of   the  differences   between   light-   and  heavy-duty
 applications   that  must  be  considered  in   the  design   of   a
 heavy-duty trap  oxidizer.   Comments  from   the  manufacturers
 restated   these   differences   (engine  size   and   load   factor,
 operating  conditions  and  temperatures,   durability,   and  ash
 accumulation),  adding   very   little  to  what   was  previously
 reported  in  the draft analysis.  The design  efforts continue  to
 be  directed  towards  a  suitable regeneration system  that  can
 handle  the   increased  exhaust  flow of  the  heavy-duty  engine
 environment  and the  generally  lower exhaust  temperatures of  a
 turbocharged  engine.   The  commenters  believe that  the  greatest
 design  challenge is  the  required   durability of   a  heavy-duty
 trap  as opposed to a  light-duty trap.

- --   While no one found  fault with the  Agency'.s identification
 of  these  design obstacles in  the adapti-on of  trap  technology  to
 heavy-duty use,  some of  the  manufacturers   strongly  disagreed
 with  the  Agency's  conclusions  that  these  obstacles  are  not
 insurmountable and traps  would  be  technically  feasible  by  the
 proposed  model year  (1990).   However, none   presented  spec.ific
, data  or engineering  analysis  to demonstrate  a LD/HD difference
 to  be an  insurmountable obstacle.   The  views of the NPRM were
 further  reinforced  by  a  document  prepared   for  the Agency, by
 Energy    and   Resource   Consultants   (ERC),   an   independent
 contractor.[6]    This  report   concluded   that  light-duty  trap
 technology can be  adapted  to  heavy-duty  use  with  additional
 development  time beyond the  effective light-duty  trap  standard
 date; line-haul  trucks  require  an extra  3-4 years,  as  their
 operating conditions  are  the   most  dissimilar  to  light-duty
 conditions,  and the  light  heavy-duty  vehicles   whose  operating
 conditions   are   more  closely  related,   require  only  1-2
 additional  years  at   the  most.     Thus,    the   extrapolation
 contained  in  the  NPRM should be  retained.    Manufacturers'
 heavy-duty test data  are  examined  in  the  following  heavy-duty
 trap  status section.

      c.    Heavy-Duty Trap  Status

      The   heavy-duty  engine   manufacturers'  trap  development
 results  examined in  the NPRM  were obtained  from  comments  the
 manufacturers submitted to EPA  in 1982  following  the  initial
 1981  particulate  NPRM  (46  FR  1910)   and  also   from  ensuing
 meetings   between  representatives of  HDD manufacturers  and  EPA
 staff.   The   latest comments received  in  response to  the   recent
 NOx/particulate  NPRM  added very little  test  data  to  what  was
 evaluated in  the NPRM.   The   following  paragraphs  will   review
 and  examine  the  current status  of  heavy-duty trap development
 work  as  reported  by  the  manufacturers.


     GM  submitted  a  summary of  trap  development  and  testing
performed  from  1981  through 1983 on  HDDEs,  much of  which  had
previously  been  submitted  to  the  Agency  for  review.   The
successful accumulation of  an  additional  70.000 kilometers on a
dump  truck  equipped  with  a  one-piece  monolith  trap  on  a
4-stroke  turbocharged  8.2L diesel engine was  the only new HDDE
testing  information  received  from  GM.   While  GM's  statement
that the  accumulated mileage (80,500 Km total)  is  short  of  the
expected  service life  of  this type  of vehicle  and  the  driving
cycle  followed  was  not  representative of actual conditions  is
correct,  trap feasibility  in some future year does  not  require
that traps be fully developed today.  In this light, the Agency
views this latest  test result as extremely promising.  At this
stage  in  the design  of  traps,  failures are  expected;  there is
sufficient time  to  work  out  trap  durability  and  regeneration
control  problems.    Despite  this,   GM  feels  that  traps  are
infeasible for  production  release  for the  1990 or  1991  model
year; GM  refuses  to  commit  itself  to  the  feasibility of traps
in the forseeable future.

     Other manufacturers commented  on the feasibility  of traps
based on  experience in their heavy-duty  trap programs.   Due to
their laboratory and field testing  results,  during  the last  two
years,   International  Harvester-  is  quite pessimistic  about  the
feasibility  of  traps,  with  durability being  the  main  design
problem.   Field experience,  to  date,  has  involved  durability
testing with three  types  of traps  on  a  6.9L  light  heavy-duty
engine.   Yet despite  failures  due  to inadequate regeneration,
IHC is willing to work towards a  trap  standard in the 1991/1992
time frame.   Mack  also is not confident  that the durability of
trap   systems  will  be   assured.    Although   Mack   expects
regeneration  and  its  control to be  feasible,  trap  durability
remains too  much of an unanswered question for  Mack  to  state a
definitive view  on  trap feasibility.   Current  work is aimed at
accomplishing  regeneration   in   actual  vehicle  use;  initial
results   produced  over  6,000 miles  of  successful  operation.
Caterpillar  believes that  trap  technology may  not  be available
for  production  by   the   proposed   1990  model  year;  however,
Caterpillar  did  not  mention a  feasible  implementation  date
beyond 1990.

     As  indicated  in its  comments,  Cummins is at an early stage
of  heavy-duty  trap  development.   If EPA   commits  itself  to
reassessing  the technical  feasibility  of  a trap  standard by
December  31, 1987,  Cummins  would feel comfortable with  a 1992
particulate  standard  of 0.25  g/BHP-hr.   However, Cummins added
the  caveat  that   it  does   not  envision  traps  by   1992.   Even
though  Volvo White  considers  current  HD trap  technology to be
virtually  non-existent,  it  believes  that trap  'technology will
be  available and  qualified by 1991  (as will  be  discussed  below,
this is conditional  on the control of  sulfur  in  diesel fuel).


     Daimler-Benz,  with  the  furthest developed  heavy-duty trap
program,  was the sole  HDE  manufacturer  to agree with the NPRM's
proposed 0.25 g/BHP-hr  trap-based particulate standard  date  of
1990 model  year implementation.   (This  also is  conditional  on
fuel sulfur  control,   in  addition to  an  allowable  maintenance
condition discussed below.)   As  described  in  its 1982 comments
to the  Agency,  Daimler-Benz is  concentrating  on  the  development
of  a  trap made  of wound  ceramic fiber.   The  latest  comments
indicate that considerable development   progress  has  been made
in  the  last  two   years.    Still,  much  development'  remains,
including the  optimization of  trap  design and  increasing  the
trap  durability  through  an   optimized  regeneration  system.
Current results of  urban bus applications  of  the traps  show a
minimum  trap service   life  of   100,000  miles,   and  a  maximum
service   life   of   less   than   150,000  miles.    With  these
encouraging   test results at  this stage  in  the  design of  traps,
the  Agency   sees  no   reason  that  the  allowable  maintenance
interval  of  150,000   miles   is  not  feasible   for  traps,  as
Daimler-Benz  indicated  in  its  comments.    As  EPA  has   stated
previously in relation  to  trap  feasibility, there is sufficient
time to work out trap  durability problems.

     The Manufacturers of  Emission Controls Association (MECA),
whose  member companies  are  supplying  the  trap  materials being
tested  by   the  HDDE   manufacturers,   strongly  supported  the
feasibility   of trap-based  standards.  Although recognizing that
development   work   remains,   MECA  stated   that   a   trap-based
standard  is  achievable, citing worldwide  test  and  development
work by its  member  companies.

     Overall, progress  in  heavy-duty trap development  has not
matched  that in the  light-duty  area over the  past two  years.
Much of this difference, however, can be attributed  to the lack
of  a  firm  target,  which  can  only  be   a  promulgated standard.
While  significant  steps still  need  to  be accomplished  in the
heavy-duty  area,   the  finding  that  light-duty  trap technology
can be  extrapolated to  heavy-duty engines and  thus,  traps are
feasible  for  future   heavy-duty usage,   remains  essentially
unchallenged.  The  key issue is  actually  leadtime,  which will
be addressed further below.

     One  issue  not considered  in  the NPRM,  and which should be
addressed here  was that raised  by  Daimler-Benz, Volvo  White,
and  several  other   manufacturers concerning  diesel  fuel  sulfur
content   and  its   relationship  to   the   heavy-duty    engine
environment.     Daimler-Benz    was    very    concerned     about
trap-plugging by non-regeneratable particulate  matter;  using  a
European  diesel  fuel,  Daimler-Benz  found  an  average  of  35
percent  of  the  non-carbon deposits  left after  regeneration to
be  sulfates.   While none  of  the other  commenters discussed  trie


issue  of   trap-plugging   by  sulfates,   Caterpillar   and   IHC
expressed  concern  that high  sulfate emissions  resulting from
high  sulfur fuel will  make  up a  significant portion  of EPA's
proposed  0.25  g/BHP-hr  standard  and possibly  exceed  it.   A
reduction  in  the sulfur content of diesel  fuel  was recommended
by  Daimler-Benz,  IHC,  Mack,  and  Volvo White;  failing  this,  or
as  an  interim  step,  the  manufacturers   recommended   that   EPA
should  adopt  a correction  factor as  part  of  its particulate
test  to  account for   the  sulfur  portion  of the particulate

     The  data  available   are  not  sufficient to  allow  a full
analysis of this issue  at  this  time.   Not enough is known about
the  Daimler-Benz trap  to  understand  why  sulfate  plugging is a
_problem there  and not elsewhere and what,  if any,  solutions  are
possible  short of  reducing  the sulfur content  of diesel fuel.
The   comments   of   other  manufacturers   presumably  apply  to
catalyst   substrate  traps,  which  generally showed   the  same
problem   on   light-duty   diesels.    (Mercedes*    trap   is   the
exception  to  this.)   As  this  is  not the  only  trap design,  or
even  that  believed to be  the most  feasible (which is  generally
thought   to  be  burner  or  fuel  additive  regenerated),   its
elimination    from   consideration   may   not   affect    overall
feasibility.   Also,  while limited areas  in California  require
low   sulfur  diesel  fuel,   the   cost  of  such  control  on a
nationwide  basis   has  not  been   determined and  would  require
significant study.

      Given  the  uncertainty  in  the  relationship  between this
issue and  feasibility,  it should  not  preclude implementation of
any   trap-based standard.    However,   the  Agency  is   open  to
further  discussion  in  this  area  and  will,  on  its  own,  be
analyzing  the  cost of controlling  the  sulfur content  of  diesel
fuel  in the future.

      Many  of  the  comments  on the sulfur  issue  addressed  the
measurement  of water,  absorbed  on the sulfate,  as particulate
emissions.   Their  concerns  center  on the  fact  that it  is very
difficult  to  reduce  the  current  conversion  of  gaseous  sulfur
dioxide  to   sulfate   (which  is   only 2-4  percent).    As   the
particulate  standard becomes more stringent, this  sulfate,with
its  water,comprises  more  and more of the  allowable emissions.
EPA  is currently   examining  a number of  different approaches
which can be  incorporated into the  test  procedure to  minimize
the   measurement  of  water.   Although commenters  recommend a
correction  factor   added   to   test  procedure  to  counter   the
problem,  the  time  constraints on  this  rulemaking  did not allow
sufficient  time  to determine  the optimum  approach.   Thus,  no
such  revisions  in  the  test  procedures  will   be made here;
potential  changes  will  be  addressed  in  a  later workshop  and


further  study  of  this   issue..   With  respect  to  the  sulfate
itself,  it  should  be   measured   as  part  of  the  particulate
emitted  as  it  is  definitely   inhalable   and   affects  human
health.  As the typical  conversion of  sulfur dioxide to sulfate
has  been  occurring  in  all  previous   heavy-duty  particulate
measurements,    and   thus,   estimates   of   trap   efficiency,
feasibility is  not affected.   Feasibility is only  an  issue when
sulfate  is significantly  increased by  a  catalyst,  which  was
discussed above.

     Some concern was also expressed regarding  the fuel economy
effects  of  trap-oxidizer  use.   A  trap  fuel  economy  penalty
incorporates  the  fuel   economy  losses   that  result  from  an
increase   in   backpressure   and   also    the   increased   fuel
consumption attributed  to  the energy  requirements of  positive
regeneration.    The  NPRM  analysis cited   a  two  percent  fuel
consumption  penalty  as  the  worst  penalty   which   would  be
observed with HD traps.

     Ford  argued that the  implementation of traps to HDDEs will
cause  fuel consumption  to  increase  by about  three  percent,
approximately  two  percent  of which  is  due  to  the backpressure
portion of the penalty (the average of the  "clean trap" penalty
of  about  one  percent  and the  "loaded  trap"  penalty  of  about
three percent).  A  second  manufacturer,  Cummins,  calculated an
approximate fuel  economy penalty of 2.6 percent  for  a 60-liter
trap; this value was  not based on actual testing.   As  repor'ted
in  the NPRM  analysis,   1.6  percent  of   the Cummins  estimated
penalty is  due to  the  increased  backpressure  and  1.0 percent
due  to  burner-initiated  regeneration  occurring  at  100-mile
intervals.  The remaining HDDE manufacturers  did  not  comment on
the  fuel  economy  effects  of  traps.   The  Department  of Energy
(DOE)  noted  a  zero  to  one  percent  fuel  economy penalty  as  a
total contribution from the backpressure on the regeneration.

     The  effect of  trap  use  on  the  HDDE  fuel  economy  is cf
course  dependent   on  the  trap system   design,   including   trap
type,   trap   size,   regeneration  type    and    frequency   of
regeneration.  Thus, it is reasonable to expect  a range of  fuel
economy penalities  for  the industry,  assuming a variety of  trap
system  designs will  be  used.    The  one  to two  percent   fuel
economy  penalty  range  documented  in  the  NPRM  analysis  is
bracketed  by the fuel economy  losses submitted  by Ford, Cummins
and  DOE,  with  the manufacturers'  values  on the  high  side  and
DOE's  range  on  the  low  side.   EPA's   own test  data  tend to
support the NPRM range.

     In  1983,  EPA tested  a  Corning ceramic  trap on  both 3  ci. s
engine and a bus chassis.[9]   The  trap caused up  to  a 2 ceroer-
fuel economy  penalty on the bus  chassis,  but caused no pen* I'./


at all  on the engine.   Steadytstate  testing  of  the engine  at
high  loads,  where the effect should be  largest,  also  showed no

     Also  in  1984,  EPA  tested  a  400 hp  HDDE  with a number  of
trap  designs.   Over  the EPA transient test,  a  Johnson-Matthey
trap mounted close to the exhaust of the  turbocharger  showed no
fuel  penalty.   A  Corning  ceramic  trap  mounted  between  the
exhaust manifold  and the  turbocharger  showed  a penalty  of  2.4
percent.   The  Corning   ceramic   traps  similarly  mounted  in
parallel showed a 9 percent fuel economy penalty.

     The  before  turbocharger  location maximizes  the  exhaust
temperature at the  trap, but also maximizes the fuel penalty as
it  directly   affects   turbocharger  effectiveness.    This   is
evidenced  by  the  fact   that  two  traps  in  parallel  cause  a
greater fuel  penalty  than a single trap.  Normally, use  of  two
traps  would  reduce  backpressure  and   reduce  any  fuel  economy
penalty effect.   However,  here  the traps  are also  acting  as
heat  sinks,  and are  removing  useful  energy that  otherwise  may
be  used by  the  turbocharger.  Two  heat sinks  are worse  than
one.   It  is  extremely unlikely that such a design would be used
on a HHDE where fuel efficiency is of  upmost importance.   Thus,
the  2.4  percent  penalty,  generated by  this  research  program
aimed   primarily   at  identifying  conditions  of  spontaneous
regeneration,  can be  taken as a definite  upper  limit  of  any
fuel penalty.

     Also  of  importance  is trap size;  an increase  in  trap  size
would  reduce  backpressure  and  reduce  the  fuel economy penalty.
In  costing trap systems  in Chapter  3,  larger trap sizes  were
used  than  those  used in the NPRM analysis or  those used  in the
EPA  tests above.   This  was  done recognizing  that even  a  0.5
percent  decrease  in  fuel  economy  penalty  would  overwhelm  the
added cost of the larger trap.

     Even  so, the trap size projected in Chapter  3 for  HHDEs is
not as  large  as  the 60-liter trap used by Cummins on its 270 hp
engine,  which  showed   the  1.6  percent  backpressure  related
penalty.   This  sizeable  penalty from such  a  large  trap  is
somewhat of an anomaly.  For example, data supplied by  GM for a
much  smaller  trap,  even  accounting for the fact that the engine
was  smaller,  showed backpressure  levels  one-third  to  one-half
lower   than   those   resulting  from  the  Cummins  trap.    This
discrepancy   may  be   due  to   trap   location-   or   operating
conditions.   The  GM  backpressure  levels   result from  actual
vehicle  road  tests,  while  Cummins'  value  was not  from  actual
condition  testing.   Thus,  the Cummins  trap seems  to have caused
an  unusually  high   fuel  economy   effect  and  a  .5-1.0  percent
backpressure  fuel  penalty  range based  on  GM's  data  is  not


      with  respect to  regeneration,  it is  possible  to  directly
 estimate  the  fuel penalty associated with  use of  a  burner  based
 system.   The  burner  us.ed   in  the  EPA  bus  testing  and   that
 forming  the  basis  for the  burner  cost  estimate made  by  Jack
 Faucett Associates were rated at  100,000  btu  per  hour.

      Using   an   estimated   burn   time   of   5   minutes    per
 regeneration,  regeneration  frequencies  of  100  miles   (used  in
 Cummins test)  and 175  miles  (maximum of 145-175 mile range  used
 in GM test),  and HDDV  fuel economies taken  from  the MOBILES
 conversion  factor analysis, [10]  the burner related  fuel penalty
 ranges  between  0.2  and  0.5  percent.   This  is  much lower  than
 the one percent  penalty estimated by Cummins.

      Thus,  overall   a  1-1.5 percent fuel  penalty would appear
 reasonable   for  a   burner  based  ceramic  , trap   system   or
 approximately  0.5   percent  attributed   to   the  burner   and
 approximately  1  percent   attributed  to   trap  backpressure.
 However,  a  fuel  additive based  trap system  would not  have  the
 fuel penalty  associated  with  the  burner.   This  type of system
 now appears to be among  the  most  promising.   Thus,  a  range  of
 0.5-1 percent fuel penalty will be used.

      Commenters   also  addressed  the potential  safety  problems
 associated  with trap usage.  The American Trucking  Association
 stated  .that   the high  temperatures  required  for   uncatalyzed
 oxidation  of  accumulated particles and  the possible  dangerous
 emissions  from catalyzed traps  are obstacles  in the design  of
.safe trap  oxidizers.  Although Cummins did  not  detail  its  trap
 safety  concerns,  Cummins did  state  that  significant  work  is
 needed in the safety area prior to the implementation of traps.
      EFA does not dispute that  the  use of trap oxidizers  poses
 potential   safety  problems.   But   the   Agency   believes   that
 through careful design  of  the  trap  system the associated  risk
 can be  reduced  to  manageable  levels.   One example  of  a  safety
 design  is   to   monitor   the  trap   temperature   to   control
 regeneration.   For  a burner system,  flame sensors can shut down
 the fuel flow if necessary.   The  trap design costed  in Chapter
 3  includes  a  number .of  such sensors.   As  for  the danger  of
 toxic  emissions  from  catalyzed   traps,  we   assume  ATA   is
 referring  to sulfate emissions,  which are a recognized problem
 with  catalyzed  traps.   EPA  is  not  aware  of   any  hazardous
 emissions   from  non-catalyzed  traps,   except   possibly   for
 catalyzing  fuel  additives,  which  would  only  be  introduced  by
 the  engine  manufacturer  if  safe.    It   is  true  that  work  is
 needed in  the safety area  as traps  are developed.   However,  the
 two production  or  production-ready  LDD  trap systems  appear  to
 be quite safe and  no  HD/LD differences  appear to  prevent  such
 safe design  of  HDD  traps.


     d.    Emission  Levels

     This   section  deals  with   comments  on  the   trap-based
 standard   of   0.25   g/BHP-hr,   separating  the  issue  of   trap
 feasibility,   as   examined  above,   from  engine-out   and   trap
 deterioration   and   trap   efficiency,   as  examined  here.   "The
 emission  levels specific for  the more  stringent 1991  model  year
 bus  and  1994  model year  HDE  0.10  g/BHP-hr  standard are  also

     EPA's determination of the design target  level generated  a
 great  deal of  comment from  the manufacturers.  Commenters  were
 critical   of   the   values   used   in  the   analysis   for   the
 deterioration   factor  of  engine-out  particulate  emissions  and
 also   the  deterioration  factor  of   the   trap-oxidizer.    The
 comments   on   the   deterioration   factor   of  the   engine-out
 particulate  emissions  and also the AQL  adjustment factor  were
 addressed in a previous  section and will  not be repeated  here.

     Several   commenters  disagreed with  the  Agency's position
 that   there  is  no   significant  deterioration of  particulate
 emissions with the use of  a  trap.   They  claimed  that traps do
 deteriorate    and   thus,   a    multiplicative   DF   of  1.0  is
 unrealistic.    Reasons  for   trap  deterioration,   according to
 Ford,  include: micro-cracks   resulting from thermal  stress  and
 high   temperatures,   leakage  at   the   trap   end  seals  due to
 warpage,   an   increase  in  the  soluble   organic   fraction  and
 regeneration   control   system  deterioration.    In   the   NPRM
 analysis,  EPA did  not  explicitly  consider  the  occurrence of
 micro-cracks,   leaks,  or  an  increase  in  the  .soluble  organic
 fraction.   All  are  theoretically  possible,  but  there  are no
 data  to  support  their likelihood;  the present durability  data
 show  no deterioration.  Therefore, a  trap deterioration  of zero
'is not  unrealistic  at  this  time.   However,   even  if   trap
 deterioration   were   a   factor   of   1.2,    it  would   affect
 feasibility;   it would only require traps to  be  applied  to an
 additional 3-10 percent of  the fleet,  depending on the standard
 level  and model year being considered.

     The   subject of  trap efficiency,   the most variable  factor
 effecting  the  emission  level,  was  also  addressed  in  the
 comments.  One commenter (Ford) did not  believe that  light-duty
 truck   trap   efficiencies   necessarily   apply   to   HDE   trap
 efficiencies,  although it presented no analysis to  support this
 opinion.     Another    commenter   (Cummins)    brought    up    the
 possibility that  trap efficiency 'depends on  the  driving  cycle
 and also  the   type  of particulate  matter;  data  indicated that
 trapping   efficiencies  for the soluble  fractions  are about  20
 percent  less   than   for  the  dry   particulates  in  a  ceramic
 monolith  trap.


       The NPRM  analysis  cited  an  efficiency  range  from  70-90
  percent  for  the  ceramic  wall-f-low  monolith  trap  and a  range
  from  50-80  percent  for  the   wire  mesh  trap's   collection
  efficiency.  Daimler-Benz's  test  results  show  the  collection
  efficiency  for  its ceramic  fiber wound trap  increasing  with  the
  filter  loading  regardless  of  the  initial trapping  efficiency.
  (An unloaded trap with a 60 collection efficiency,  increased to
  80  percent  efficiency  with   15  percent  loading,  90  percent
  efficiency  with  40  percent loading,  and  97  percent  efficiency
  with 70  percent  loading.)  GM  commented  that it  hasn't seen  the
  efficiencies that EPA reported out  of  its traps.   However,  at
  another  point  in its  submittal,  GM  stated  that  "in  spite  of
  repeated structural  failures, with  ceramic  monoliths,  we  have
  continued   their  development   because of  their  high  trapping
  efficiency  and overall potential once the  control  problems  for
  a consistent regeneration  are  resolved."[11]        ..      	

    .   Based  upon  the  above, an 80 percent  efficiency was  chosen
  in the NPRM to  represent an obtainable .feasible  trap efficiency
  level in the 1990 timeframe.   One  commenter  disagreed with this
  efficiency   level  referring to  testing of  a  trap-equipped  bus
  engine  conducted by  Southwest  Research  Institute   (SwRI)  for
  EPA.[9]   The transient  test  particulate emissions  of  the  DDAD
  6V-71 engine were reduced  61  percent using  a  ceramic trap over
  -the  FTP; over   a  bus  cycle,  total  particulate  was  reduced  68
  percent.  This  testing  was done on an old engine  notorious  for
  a  high  soluble  organic   fraction  of  its  particulate,   which
  explains the  low collection  efficiencies.  Current technology
  engines  have much lower HC emissions and  lower  soluble organic
•vl fractions   (SOF)  which  should  result  in  a  much   higher  trap
  efficiency   as   indicated   in  Cummins'   comments   that   trap
  efficiency   increases   as   the  SOF  decreases.   Other  testing
  conducced  by SwRI[12]  did  result in a  higher  trap  efficiency; a
  Cummins   NTC-400  engine   equipped   with   a   Corning  trap  was
  effective   at  reducing  particulate   emissions  by  85  percent.
  Thus, an   80  percent  efficient  trap  is  still  reasonable  with
  respect  to  a 0.25 g/BHP-hr standard,  if not on  the  low side of
  what traps' actual collection efficiency will be.

       Applying  the  trap  deterioration  factor,  SEA  adjustment
-  factor and  the  trap efficiency to  the engine-out  target  level,
  (0.42-0.54   g/BHP-hr  from  above),  yields  an emission  level of
  0.10-0.13  g/BHP-hr.   Thus,  at  the  0.25  g/BHP-hr  standard,  traps
  will  not   be  required  on  all  engines;  the  technically -most
  difficult   applications  will  be  able  to be  excluded  from trap
  usage,  which   is  desirable   given  the  new  nature  of  this
  technology.  With averaging,   approximately  70  percent of  the
  HDDEs  will  be  trap-equipped  in   order   to  meet  the  0.25
  g/BHP-hr.    The  percentage  of  the  fleet  requiring  traps  should
  decrease to  approximately  60  percent by 1994,  as the engine-out
  target  level  decreases to 0.42 (discussed  in  Section above);
  this assumes a trap efficiency of 85 percent.


      Limiting  traps to only highly  efficient,  ceramic  wall-flow
 monoliths,  even  lower  levels  can be  achieved.   Thus,   assuming
 85-90 percent  efficient  traps,  the  engine-out  target   level  of
 0.42-0.54  g/BHP-hr  results  in  a  emission  level  of  0.05-0.08
 g/BHP-hr.which would comply  with the 0.10  g/BHP-hr  standard.

      By  1994,  the engine-out  target level  is  projected to  be
 0.42   g/BHP-hr.    Assuming  unchanged  deterioration   and   SEA
 adjustment  factors and 90 percent  efficient  traps,  this results
 in an  emission  level of   0.05-0.06  g/BHP-hr.   Under  a  0.10
 g/BHP-hr  standard  and with  averaging (excluding urban  buses),
 roughly  90  percent of the HDDEs will be  trap  equipped.

      In  commenting on  stringent  particulate  emission standards,
 in addition  to trap  technology,  many commenters addressed  the
 use of methanol  fuel  in  diesel engines  as a method  for  further
 reductions   of  HDDE   particulate   emissions.    Views   on  this
 subject  were widely  held.   NRDC and other environmental  groups
 believed  that  EPA  should  establish both  NOx  and  particulate
 standards  based  upon  the  use  of  methanol  as  a   fuel  in  new
 engines  and  also set  regulations  to assure  the existence  of a
 supply  and  distribution   for  methanol   to   fuel  heavy-duty

      Comments  from HDDE  manufacturers  expressed  caution  over
 the use of methanol  fuels.   Saab-Scania stated that it  was  not
 prepared  to provide  methanol-fueled  engines  in  transit buses in
 1990   due   primarily  to   the  uncertainty  of  the  unregulated
 pollutants  and  their  health  effects.   This  comment was  fairly
 typical  of  those by other manufacturers addressing  this issue;
'concern over the  technological  aspects  of methanol-fueled HDDEs
 was  a  minor  issue compared  to the  potential  health   risks  of

      New Jersey Transit  and  other  public  transit  authorities
 believed that  EPA  should  analyze  and  further  evaluate  the
 feasibility  of  methanol  as   an alternative  fuel,  expressing
 concern   about    the   difficulties   related    to   storage,
 distribution,  operating  range limitations for  vehicles and the
 risks of  formaldehyde emissions.

      While   EPA  continues  to   believe   in   the  potential  of
 methanol  in  this  area,  it considers  it   premature  to actually
 set  standards  requiring  the  use  of   methanol.    Many  basic
 questions  remain  to  be dealt  with before  widespread adoption of
 methanol  fuel will be  possible.   Therefore,  while  continuing to
 encourage  the  development  of  methanol-based  technology,  EPA is
 taking no  action at this  time on methanol-based standards.


     e.    Leadtime
           I_B^^^_WM_«^_^«««            • —

     i.    0.25 q/BHP-hr Standard

     In its NPRM analysis of  leadtime,  EPA concluded that  there
 appeared  to  be sufficient  time for the manufacturers to design,
 develop,  and  prepare  trap  oxidizers for 1990  model year HDDEs.
 In  their  comments,  all  the manufacturers,  some  to  a greater
 degree  than  others,  were  cautious  in predicting  a  date  for
 traps  to  be  in  production  on  HDDEs.   Only  one '  commenter
 (Daimler-Benz) agreed  with  the Agency's proposed implementation
 date,  albeit  conditionally,  as  discussed  above.   The   other
 manufacturers  disputed  EPA's  analysis   that  traps   would   be
 feasible  for  1990  model year application.  In  their submittals,
 the majority of the manufacturers  did provide  alternative  dates
 to  the proposed 1990  model  year effective date.   As  reviewed,
 International  Harvester  and  Cummins  indicated  a  willingness  to
.work   towards   a   trap-based  standard  in   the  1991   to  1992
 timeframe.  Volvo White expressed  its  belief  that traps will  be
 available  and qualified  by  1991.   Ford  did not believe  traps
 could  be  implemented  prior  to  the  1991  model year,  if  then.
 The  remaining manufacturers  of  HDDEs were not certain at what
 date in the future traps would be available.

     In    re-examining   the   necessary    leadtime   for   the
 implementation of   a  trap-based   standard,  it  appears   the
 effective  date should be  delayed  from  the  proposed 1990  model
 year  to  the 1991 model year  for several  reasons.  First,  there
 has  been  little  apparent   progress  in  HD  trap  technology
 development  by the heavy-duty industry over  the past  two  years
 (Daimler-Benz  being  the  most  notable  exception).  Light-duty
 trap   technology  has  continued  to   progress  and  insofar   as
 heavy-duty trap   technology  is  an  outgrowth  of  light-duty
 technology,  heavy-duty technology  has  progressed  even  without
 any overt  effort by HD  manufacturers.   However, not all of  this
 lack  of  progress  over the past  two  years is recoverable and an
 extra  year of  leadtime would  appear  reasonable.

     Second, "the  promulgation of  these  standards,  is  somewhat
 later  than  originally anticipated.    (March  of  1985  vs.   late
 1984).  While  not constituting an  entire year, the  leadtime  was
 tight  to  begin with  and an  extra year  is  reasonable   for  this
 reason as  well.

     Third,  the Clean Air Act requires that  revised heavy-duty
 HC,  CO  and  NOx   be   at  least  three  years  apart.   While  not
 applying   directly  to  this  particulate  standard,  it  appears
 reasonable  to follow  this   approach  in this   case.  The  5
 g/BHP-hr  NOx  standard  is  being  implemented  in 1991 and  it  is
 reasonable to have the  particulate  standard change at the same


      ii.    0.10  q/BHP-hr  Standard

      In  order to  comply  with the more  stringent 0.10 g/BHP-hr"
standard,  traps must  be  85-90  percent efficient  depending on
their  engine-out particulate levels.  Presently  this efficiency
cannot  be  obtained  by  all  trap  designs,  and  the  design of
high-efficiency  traps  is  generally considered to be technically
more  difficult than  lower efficiency designs.

      In  their comments on a  0.10  g/BHP-hr standard, most of the
HODE  manufacturers argued strongly  that this standard  was not
achievable.   Only  Daimler-Benz and Volvo White believed  that
the  necessary trap  efficiencies  were feasible, and  this  was in
relation  to the proposed bus  standard,  as discussed below.  The
Engine  Manufacturers Association  (EMA)  and Ford believed  that
trap  efficiency  must  be 85  percent to meet  a 0.10  g/BHP hr
standard  and believed  this  level  not  possible.   GM added  that
EPA   disregarded  the  variability   of  trap  efficiencies  in
assuming  a  90 percent efficient trap was  possible.   Aside  from
the  general comments on  trap  efficiency/  technical comments did
not  address  specific  difficulties  involved  in  meeting  a  0.10
g/BHP-hr   as  compared   to   a  0.25 g/BHP-hr  standard   (i.e.,
obstacles  in the way of obtaining  a  higher trap efficiency).

      The   majority  of   the  coramenters  .felt   this  level  of
particulate emissions  was  unobtainable for two reasons.  The
first being that the  required  trapping  efficiency  would not be
possible   by  the  1991  model  year.   The  other  reason  was
discussed  previously:  high sulfate emissions,  resulting  from
.high  sulfur fuel, will  exceed a 0.10 g/BHP-hr standard.  While
Daimler-Benz shared  the concern over the sulfur  issue, the  HDDE
'manufacturer  stated  that   the   0.10   -g/BHP-hr  standard  was
obtainable   for   1990  model  year  buses,  depending   on  an
improvement  to  the   particulate  measurement  accuracy  at   low
levels.   The  Manufacturers  of  Emission   Controls  Association
also  believed that  a  0.10 g/BHP-hr  standard was achievable  for
1990  model  year  buses.

      Trap  efficiency  may be  increased by  either  employing  a
different   trap  type  or   by  making  design  changes to  a  lower
efficiency   trap.   Of  the  trap  designs  currently   considered
promising,  the  ceramic   monolith  trap   is  the  most efficient  -
its   efficiency  can  be   above  90   percent.   A ceramic   trap
efficiency  is related  to the porosity  of its honeycomb  matrix;
high   porosity  results   in  low  efficiency  and  vice   versa.
Engineering challenges that  result  from a decrease in the  trap
porosity   (increase   in  trap   efficiency)    include   faster
backpressure rises,  which must be compensated  by increasing the
trap   size   or  by  more  frequent   regeneration.   The   forrer
solution may  also  solve  the  potential  increase  in ash  or  fuel


additive accumulation.  The  latter  may not.   A  larger  trap can
create additional  design  problems  itself  (greater  stresses and
heating requirements for regeneration)  as  were  discussed in the
NPRM  analysis  of   light-duty/heavy-duty  design  differences.
More frequent regeneration  is  fairly simple for  a  burner-based
system,  but  may  be  much  more  difficult  for  catalyzed  or
fuel-additive   based   systems   where   naturally   occurring
temperatures are relied upon to induce regeneration.

     Due   to   the   increased   difficulty   in   designing   a
higher-efficiency  trap  capable   of  complying  with   a  0.10
g/BHP-hr standard,  the technical feasibility of  all  1991 model
year  HDDEs  complying  with  this  standard is  not  likely. "  By
establishing a  1994  0.10  g/BHP-hr standard,  the Agency believes
that the additional  three  years  will allow for  the development
of   higher   efficiency  traps   with  more  time   to  optimize
performance and durability while minimizing cost.

""Theoverall  ""difficulty    of  .achieving   these   high
efficiencies also  depends  on  the number  of engines  needing  to
be  so  equipped.  It  is  likely  that a  number  of  trap  systems
employed to  meet  the  0.25  g/BHP-hr averaging  standard will  be
85-90 percent efficient.  Others  will  be less  efficient.  Thus,
for  a  few  engine  urban  transit  buses,  for  example,  a  '0.10
g/BHP-hr standard  should  be quite  feasible.  In actuality, few
bus  engines  are  currently marketed  in the U.S.   GM  dominates
the  market  with its  6V-92TA and 8V-92TA  bus  engines.    Cummins
has  sold a  small  number of its VTB-903  engines  in  buses in the
past,  but  is   not  currently  doing  so.   A  small  number  of
foreign-based manufacturers,  such  as  M.A.N.  and Daimler-Benz,
have  recently  begun to market  bus engines  in  the U.S.  Thus,
bus  engines  represent  a  relatively  small  subset  of  HDDEs.
Developing a trap-oxidizec  system for transit bus  use  may also
be  considerably  easier  than   for  most  HDE applications.   An
EPA-sponsored  report  by  Energy  Resource Consultants,  Inc.[6]
mentions the following reasons this may  be so:

     1.    Durability  and  reliability  requirements  would not be
nearly as strict as for most other  types of heavy-duty  vehicles.

     2.    Buses have  a rather  predictable operating  cycle, and
and  one  which  includes   a great  deal  of acceleration.   The
frequent occurrence  of moderate  high exhaust temperatures  as  a
result would help to make a self-regenerating system  feasible.

     3.    Transit  buses   universally   receive  regular   service,
often on a daily basis.


     Thus,  at  minimum,  bus engines should be  no  more difficult
to  trap-equip  than other  HDDEs  and,  at  best,  could  be much
easier   to   trap-equip.    Coupled  with   their   small  number,
developing   high-efficiency  traps  for  bus   engines   should  be
feasible  at the  same  time as traps  are generally  employed  on
HDDEs, or 1991.

     Equipping   all  HDDEs  with   high-efficiency   traps  will
require  additional  time  beyond 1991._  Having  buses  operating
with  such  traps  will certainly  provide useful  data,  but such
data cannot be employed  any  sooner than three  years  after the
buses  begin service, since time is required to obtain the data
and  design  and  tooling  must  also be  performed.  Of  more use
will  be durability data  generated on prototype  non-bus HDDVs
equipped  with   high-efficiency   traps   after  the  bus   engine
designs   have   been  set,  but   prior   to   bus  introduction.
Providing  only  two years between standards  can  reasonably  be
ruled  out  due  to  the  need  to apply  such  traps  to line-haul
"HDDEs,  which have  very  long  lives and  which require extensive
durability  data.   The  argument  can be  made that  three years
should  be  sufficient  to  incorporate  such durability  data.   As
this  also  coincides with the Act's requirement  for HC, CO, and
NOx  standards,  it  appears the most reasonable interval  time as

     .C.     Conclusions

     1.     Near- and Mid-term NOx  and Particulate Standards

     As  a  result  of  the proceeding analysis of the comments,
EPA  has concluded  that  the proposed standards  of  6.0 g/BHP-hr
NOx  and 0.60  g/BHP-hr  particulate are technologically feasible
and  that   the  appropriate  date  for   implementation  of   these
standards  is the 1988  model year.

     EPA has also  concluded  that   engine-out  emission  standards
of   4.0  g/BHP-hr   NOx  and  0.40  g/BHP-hr   for   HDDEs  are  not
technologically   feasible  using   any  known  emission   control
technology.  Information provided   in the comments has, however,
lead EPA   to  modify  the NPRM  analysis  and conclude  that  an
engine-out    NOx   emission   standard   of   5.0    g/BHP-hr   is
technologically   feasible  by  the  1991  model  year.   The  lowest
feasible engine-out particulate  level,  given  the  5.0 g/BHP-hr
NOx  standard,  appears  to be  0.50 g/BHP-hr.

     With  respect  to fuel economy, the  6.0/0.60 standards are
expected to  cause  a  0-2 percent  fuel  economy  penalty   in the
near term  and that  this penalty  will be  erased by  1991.  The
NOx  standard of  5.0 g/BHP-hr is expected to  cause  approximately
a 1  percent penalty  initially,  decreasing  to approximated  1/2
percent in a few years.


     2.    Trap-Based Particulate Standards

     As a result of the proceeding  analysis  of the comments  and
 additional   information,    the   Agency   concluded   that   trap
 technology    is    feasible   for   heavy-duty   diesel    engine
 application.   This conclusion was  extrapolated from the  status
 of  light-duty trap technology  and  the design  effort  necessary
 to  adapt  this technology to heavy-duty usage.  Light-duty traps
 have  been  proven  to  be  a  feasible  control  of   particulate
 emissions   from   light-duty  vehicles.    Although  'conditions
 specific   to   the   HDDE   environment    require   considerable
 development  in order to  apply  LD  trap  technology to  HD  usage,
 these  obstacles   are  not  insurmountable  and  with   adequate
 engineering  effort traps  should be a feasible  control  method of
 particulate  emissions from heavy-duty vehicles.

     EPA  has also concluded  that  the 0.25  g/BHP-hr trap-based
 particulate  standard  should be  feasible  for 1991  model  year
'HDDEs; the  0.10  g/BHP-hr  trap-based particulate standard  should
 be  feasible  for  1991 model year  urban  buses  and  for   all  1994
 model  year  HDDEs.   The analysis determined  that  the 1991  model
 year  HDDE  standard,  with averaging, will  require  80  percent
 efficient traps on roughly 70 percent  of the fleet.  This  would
 decrease  to  about 60 percent after the  initial years.   The '1991
 model  year  bus   standard will  require   the use  of nearly  90
 percent  efficient  traps on all buses.   The  1994 model  year HDDE
 standard  will require the use of 90  percent efficient   traps- on
 roughly  90 percent of all HDDEs.



     1.    "Draft  Regulatory  Impact  Analysis  and  Oxides  of
Nitrogen Pollutant Specific Study," U.S. EPA, OAR, QMS.

     2.    "Assessment    of    Domestic    Automotive    Industry
Production  Lead  Time  of  1975/76  Model  Year  -  Final Report."
Aerospace Corporation for U.S. EPA, OMSAPC, ECTD, December 1972.

     3.    "Trap-Oxidizer  Feasibility Study,"  U.S.  EPA,  OANR,
OMS, ECTD, SDSB, Public Docket No. A-83-32, February 1982, .

     4.    "An   Updated   Assessment   of   the-  Feasibility   of
Trap-Oxidizers,"  Regulatory  Support Document,  J.  Alson  and R.
Wilcox,  U.S.   EPA,  OANR,  OMS,   ECTD,  SDSB,  Public  Docket  No.
A-82-32, June  1983,

     5.    "Trap-Oxidizer  Technology  for   Light-Duty  Diesel
Vehicles:   Feasibility,  Costs  and Present  Status,"  Energy and
Resource  Consultants,   Inc.,  Final   Report   for   U.S.   EPA.
Contract NO. 68-01-6543, Public Docket No. A-82-32.

     6.    "Particulate   Control  Technology   and  Particulate
Emission  Standards  for Heavy-Duty  Diesel Engines,"  Energy  and
Resource  Consultants,  Inc.,  Report  to   U.S.  EPA,   Office  of
Policy  and  Analysis,  EPA  Contract  #68-01-6543,  December  11,

     7.    "Mercedes  to  Use  Traps on  1985  Turbodiesel," Wards
Engine Update, Volume 10, No. 13, July 1984.

     8.    "'86  VWA  Diesels  Will  Have   Traps,"  Wards  Engine
Update, Volume 10, No. 20, October  15, 1984.

     9.    "Preliminary  Particulate  Trap Tests  on  a 2-stroke
Diesel  Bus  Engine,"  Ullman,  T.  L.,  Hare,  C.   T. ,   Southwest
Research  Institute,   Baines,  T.  M. ,  EPA,   SAE  Technical Paper
Series, 840079,  February  1984.

     10.   "Heavy-Duty   Vehicle  Emission   Conversion  Factors
1962-1997"  Smith, M.C.   IV,  U.S.  EPA,  OAR,  OMS,  ECTD,  .SDSB,
EPA-AA-SDSB-4-1, August 1984.

     11.   "General  Motors'  Final  Comments  on  the  October  15,
1984  Notice  of  Proposed Rulemaking with  Respect  to Gaseous
Emissions Regulations  for 1987 and Later  Model Year  Light-Duty
Vehicles,   Light-Duty  Trucks,   and  Heavy-Duty   Engines  and
Particulate Emission  Regulations for 1987 and Later  Model  Year
Light-Duty  Diesel   Trucks   and   Heavy-Duty  Diesel   Engines,"
Submitted to  Environmental  Protection Agency,  Washington, D.C.,
December  17,  1984.

     12.   "Heavy-Duty    Engine    Exhaust    Particulate    Trap
Evaluation,"   U.S.    EPA,   OMSAPC,   ECTD,   EPA  460/3-84-008,
September 1984.

                           CHAPTER 3

                        ECONOMIC IMPACT

     This chapter  analyzes  the costs of complying  with  the new
NOx and  diesel  particulate standards  in  light of  the  comments
received  in response  to   the  NPRM.   These  comments  at  times
supported and at times disputed the EPA cost  estimates;  some of
these  comments  prompted  revisions of  the costs  given  in the
NPRM,  and are outlined below.

     The chapter begins with a synopsis of  the methodology used
in the  Draft Regulatory Impact Analysis  (RIA) to  generate .the
cost  estimates  in  the NPRM.   Following  the  synopsis   is  the
Summary  and Analysis  of Comments, which  is  divided  into three
cost sections:  LOT,  HDGE,  and HDDE.   Within each  section  is  a
summary  of  the  applicable  comments,   discussion  of   how  -the
comments compare  to the information  contained in the  NPRM,  and
any  reanalysis  as  necessary.   Each  section  closes  with  a
summary  of  the  final  cost and  cost-rel-ated values used in the
final   economic   impact   analysis.   This   is   followed  by  a
discussion  of socioeconomic impacts.

I.   Synopsis of the NPRM Analysis

     This  chapter as  originally  presented  in the  Draft  RIA
examined  the compliance costs  of the  proposed NOx  and diesel
particulate  standards  for LDTs   and  HDEs.   It   included  the
manufacturers'   fixed  costs   of  pre-production  (research,
development,   and  testing   (RD&T),   including    certification
testing),   and  their   variable costs  of   production  (emission
control  hardware  component costs),  as well  as  user costs of
increased  purchase  price,  fuel economy  losses, and maintenance
cost changes.   The  chapter  was divided into  two  sections which
discussed,   respectively,  the  actual  manufacturer  and   user
costs,  and  the  socioeconomic  impacts  of  such  costs.   The  first
section  is  the  more lengthy one,   and  received  the  main bulk of
the  comments.   It  is  summarized  below.   The   socioeconomic
impact   section,"  which  included  manufacturer,  regional,  and
national  effects  on  sales,  cash  flow,  employment,  balance of
trade,   and  consumer   prices   received  comments  on  two issues
only, and therefore need not be reviewed in full.

     Commenters on  costs  focused  on  alternative values  to the
costs  derived  by EPA  rather  than on  the  methodology used,- and
therefore  the methodology  is  described here  only  briefly.   Any
interested  parties  may consult the Draft  RIA for  more  complete
information on  the  cost  derivation methodology  and actual  cost
values which were presented in the NPRM.


      A.     Cost  to  Manufacturers

      In  EPA's   analysis,  manufacturer  costs  for  each  of  the
vehicle/engine  groups —  LOT,  HDGE,  and  HDDE  —  included  the
fixed 'costs of  RD&T  and the variable costs of  hardware.   Fixed
costs   were    determined   by   estimating   the   number    of
recalibrations,  design  modifications,  and  the amount  of  new
testing  necessary  to convert  present  systems  to  those  which
could meet the  standard.   Numbers  of  calibrations  needed  per
engine  family were  combined  with  numbers  of  engine  families
needing  the work,  estimated  hours  of  effort  per calibration,
hourly  rates  for  labor, overhead and  parts,  and  a  10  percent
'contingency  factor   to  derive  a   dollar   value  for   total
recalibrations.    Similar   estimates  were  made  of   the   time
-necessary   for  redesign  and for  completely  new, general  system
designs, such  as that required for particulate  traps.

      Testing  costs to prove mechanical  integrity  were based  on
miles of testing necessary, average  speed,  and hourly rates for
labor and  overhead;   such  test  costs  were  shared  with  other
testing  programs when  applicable.   Certification testing  costs
included the same  type  of  mileage  accumulation costs,  as  well
as fixed costs  of  $1,500 per  emission test for LDTs  and  $2000
per emission  test  for HDEs.

      For LOT,  HDGE,  and  1987  HDDE  proposed  standards,  it  was
assumed  that  these  fixed  costs  would  be incurred  in the  fcwo
years prior to  implementation of the standards;  for the  1990
HDDE standard,  four years  were  allotted  due to  the  longer
development time needed for trap-oxidizer  systems.   The sum  of
all these  costs was  apportioned  over  five model years for  LDTs,
and  three  model years  for  HDEs  (due  to introduction  of  the
second   set of  HDE  standards  after three  years) .   Costs  were
presented    in   both   undiscounted   and   discounted   forms.
Discounted costs were  calculated at a 10 percent  discount rate
to the first year  for which  the  standard was  applicable  (1987
or 1990).   These costs  were then spread over projected sales to
determine  an  average  cost per  vehicle or engine due to RD&T.

      Variable  costs to manufacturers arose from the  addition of
new  hardware  and,   in  some  cases,  credit  was taken  for  the
removal  of old  hardware components.   For  LDTs these  component
costs were developed using the  Rath and Strong  methodology[1]
as  discussed   in  the  Draft   RIA  and   include   overhead  and
manufacturer   profit.   For  HDGE and  HDDE   NOx   and  non-trap
particulate control  on HDDEs,  component  costs  were  developed
 from costs for  similar  pieces of equipment  on  current engines,
with inclusion  of  factors  for different  material  costs due  to
different   component  sizes.    Particulate  trap costs were  taken
directly from the Diesel  Particulate  Study.[2]  These component
cost estimates  were  combined  with projections  of  the technology


changes  which would  be necessary  to meet  the  new  standards,
rr.arket  snares  of  various  technology mixes  and vehicle/engine
types,  and  projected  sales,  to develop  both per vehicle/engine
and  aggregate manufacturer costs  for hardware.  Hardware  costs
were  combined with RD&T costs, with  appropriate discounting  at
10 percent, to determine total manufacturer  costs.

      B.     Cost  to Users

      Costs  to users  were  based  on increases  in  first cost  at
the   retail  price   equivalent   (RPE)   level   and   additional
operating   costs   due   to    changes   in   fuel   economy   and
maintenance.   The  first  price  increase per vehicle  includes 'the
average   hardware   cost   for   that  vehicle's  new   technology
application  and   the  per  vehicle  share   of   RD&T,   which   is
apportioned over  the three  or five  years after implementation
of the  standard  as discussed above.

      The  lifetime  cost changes per one  percent  of  fuel economy
change  were   calculated  from  fuel  price,   average  base  fuel
economy  and  lifetime mileage per  vehicle  or  engine  category,
using a  10  percent discount rate.  Overall  fuel economy changes
expected  were  estimated  in   the  Technological   Feasibility
Chapter  according  to  the  types-of technology necessary to  meet
the  standard.  Lifetime fuel economy  costs  due  to  the standard
could thus  be calculated by  multiplying cost  per one percent
change  by the amount  of  change expected.

      Maintenance   costs  were  determined from  any  additional
maintenance operations deemed necessary  for the new technology,
the  expected  number  of additional maintenance  operations  per
 lifetime,  and the cost  per  occurrence.   These  costs were  then
discounted  at  the   usual   10  percent  rsts,  from   point   of
maintenance to point  of  sale.   The costs were  then  apportioned
 two  different ways.   In  the first case, they  were  apportioned
.over  just  the  engines  requiring  the  new  technology  and  the
corresponding maintenance,  and   in  the  second  case  the  costs
were   apportioned   over  all  engines.    In  appropriate   cases
credits  were  taken  for  maintenance  which  would  be reduced,
using the  same methodology.   Maintenance costs, either negative
or  positive,  were  added to  fuel  economy costs  and  first  price
 increases to  give  the total  lifetime  user cost  of the standard.

      Aggregate costs  to  the  nation,  including those to  both
manufacturer  and user,  were  then  calculated for each  vehicle or
engine  group.  Hardware costs  plus operating costs  of fuel  and
maintenance  were  multiplied  by  number  of   vehicles  affected,
 according   to  future  sales  projections.   As  noted  above,  a
 5-year  period was  used for  LDTs  and 3-year  periods  for  HDEs.
These  values   were   discounted   to  the   proposed  year   of
 implementation,  and  added  to discounted RD&T  costs, to  yield


 the  net present  value  aggregate  cost in the year the  standards
 begin.   Costs  to  the nation  were  -also expressed in terms of  per
 vehicle investment  in  1987.   All  values  used  throughout  the
 chapter were 1984 dollars.

 II.   Summary and  Analysis o£ Comments

      A.     LPT NOx Standard

      1.     1988 NOx Standards

      a.     Cost to LPT Manufacturers

      Comments  on  the manufacturer costs attributable to  the  new
 LDT  NOx standards  were neither numerous nor lengthy,  with only
 two   manufacturers  estimating  retail  price  increases  due  .to
 hardware and  RD&T.  The  price  increases  in the comments were
 stated  without derivations,  and the  methods used  by  EPA  to
 estimate the  costs given in the NPRM were  not challenged.   The
 manufacturer and  EPA  final  costs,  however, were not  based  on
 comparable  fleets.  Manufacturer costs were based on only  those
 vehicles requiring  new  technology,  while  the EPA  final  costs
 developed  in  the NPRM were  expressed  as  an average over  every
 LDGT or LDDT,  including  those  vehicles  which will already have
 the  hardware  in  place prior to  implementation of the  standard.
 EPA  developed  an  average  per  vehicle cost for a  LDGT  requiring
 new   technology  in  the Draft  RIA,   as  part of  the  process  of
 developing  the fleet  average LDT costs.   This cost ($140)  was
 presented   in   the  Draft  RIA.   The  EPA  and manufacturer cost
'estimates  are as  follows:    ....

                   Retail  Price  Increase Per  LDT

      Draft  Analysis  $35         LDDT Average
               ,  , '   $44-87       LDGT Average
                      $140        LDGT with new technology

      Chrysler    .    $80    _  _  LDT with  new technology

      Toyota  "        $100-250    LDTt with new technology

      As can be seen,  when the same  figures  — cost  per  vehicle
 with  new   technology  —  are  compared,   those   given  by  the
 manufacturers  are  comparable to those  estimated  by EPA in  the
 NPRM.  Therefore,  EPA  sees  no  need, to alter its  analysis of  the
 LDT   component  cost values  or  RD&T  costs  which were used  to
 develop the average per vehicle values presented  in  the NPRM.

      i.    Fixed  Cost

      In developing  the costs  for  the  final rule, the original
 allocation  for RD&T costs  estimated by  EPA for this  standard
 will 'be used  here,  but  shifted one  year  to  concur  with  the


shift in the year of introduction of the  standard.   This amount
is the  sum  of the  RD&T  costs  for  LDGTs  and LDDTs,  and totals

     ii.   Variable Cost

     The   average  costs   have    been   updated   to   reflect
manufacturer comments  on projected  technology  mixes received in
response to the  NPRM.  Present  technology mixes are  taken from
1985 model  year  certification data,  which provides manufacturer
sales  projections by  engine  family  and,   hence,  by  emission
control  technology  type.  The  manufacturer data  from comments
and the  1985  certification data  provide the basis  for revision
of  the  projected  technology  mixes  and  subsequent  revision  of
the average LDGT hardware  costs,  which  are calculated  in  the
same manner as used in  the  Draft RIA.   The projected mixes were
in  most  cases confidential  on   a   manufacturer-specific  basis,
and are included here  only  in  general form.

     The   new   present    (1985)   and   projected   technology
application mixes  according  to   LDGT engine size  are  given  in
Table  3-1.   The  same  information  for  all LDGTs  combined  is
presented   in  Table  3-2.    The  1985  model  year  mixes  were
converted to  projected  1987 model  year  (pre-standard) mixes by
applying  technology  changes  on  specific   engine  families  as
indicated  by  manufacturers   in  confidential   comments   to  the
NPRM.  As noted in the NPRM,  the comments  verify that even with
no  increases  in  the stringency  of  the NOx  standard,  there is  a
clear   trend   away  from   oxidation   catalyst   systems   to
three-way-catalyst systems on LDTs,  apparently for reasons such
as improved driveability and fuel  economy.   The trend for LDGTs
is  predominantly  toward   three-way  closed-loop   systems   as
opposed   to  three-way   open—loop   systems.    Therefore,  this
analysis projected  that  for the  1988 model year (the  first year
of  the  new  standard),  all  remaining oxidation  catalyst  systems
and  all three-way  open-loop  systems will  convert  to three-way
closed-loop technology.

   -  The  overall hardware cost  for  each vehicle  undergoing  a
technology  change,  according   to   number  of  cylinders  in  the
vehicle,  is given in Table 3-3, which  is  a summary  of Tables
3-4  through  3-6  in  the  Draft  RIA.   Costs   are  derived  by
subtracting  the  costs of hardware components   removed  from  the
costs  of hardware  components  added;  estimates of  these costs
were   originally   determined   using   the  Rath    and   Strong
methodology,[1]  and are unchanged  from the Draft RIA.

     The  hardware costs  as  presented  in  the Draft RIA  and the
projected change  in the  technology  mix,  as discussed  above,  are
combined  to give  an  average  hardware  cost  upon  implementation
of  the  1988  model  year standard  for  LDGTiS  and  for LDGT2s
as  follows:

                                     3-6 -
                                  Table 3-L

                          Light-Duty Gasoline Trucks
            Percent Technology Usage 3y Model Year and Engine Size

 1985 MY*

"6-cylinder ~ '-"I. ~~

 1987 MY projected**
 1988 MY projected**

 4-cylinder   .  •    0
 6-cylinder   •      £
 ALL            . •   0

 1985 MY*

 1987 MY projected**
 1988 MY projected**

 6-cylinder    .     0
 3-cylinder         £
 ALL         .       0
Based on manufacturers' confidential sales projections'.
Based on manufacturers' confidential comments to the NPR4.


                             Table 3-2

                     Ligr.t-Duty Gasoline Trucks
       Percent Technology Usage by Model Years  for All LDGTs
                                    Percent Technology Used
Model Year
1983 and

 '   4.2
•   0.0


Based  on   confidential   sales  projections   provided  by
manufacturers as part of the certification process.
Projected based  on  confidential  manufacturer comments  to
the NPRM.
Also   includes    three-way   closed-loop   plus   oxidation
catalyst systems (see Table 3-1).


                           Taole 3-3
                    Lignt Duty Gasoline Truck
                Emission Control System Hardware
              Cost Per Vehicle with New Technology
  	Technology Change	    	Engine Size	
  From      	To	     4CYL       6CYL       8CYL
Oxidation   Three-way Closed Loop     $106    •   $133
Three-way   Three-way Closed Loop


      It is projected  that 48 percent  of all  LDGT-s  will
      require  technology changes  between 1987 and 1988.

      First,    4-cylinder    LDGTiS,     representing    21.2
      percent   of  all   LDGTt   sales,   are  converted  from
      oxidation   catalysts   to   three-way   closed-loop
      catalysts.   At $106  per conversion,  the contribution
      of 4-cylinder  engines  to   the  LDGT,  hardware  cost
      is:   0.212  X $106 - $22.47

      Six  cylinder  LDGTts,  representing  26.8  percent  of
      all  LDGTis,   are  changed  from  oxidation  catalysts
      to  three-way  closed-loop   catalysts,   at  $133  "per
      vehicle.    The  6-cylinder  contribution to  the  LDGTt
      hardware  cost is  then:   0.268 X $133 = $35.64

      The  remainder  of  the  LDGTi market,  52  percent,  is
      projected  to  already  have  the   required  technology
      (38.5 percent three-way  closed-loop  catalysts  and
      13.5  percent  three-way  closed-loop   plus  oxidation
      catalysts)   in  place for  the  1987  model year.   (In
      fact, most of these vehicles, 49.4  percent  of the
      market,   already  have the hardware  on  the 1985 model
      year  vehicles.)    No. costs  are   incurred  for  these

      The  average hardware  cost  per   LDGTi  in  the  fleet
      is the   sum  of   the contributions  by  the  4-  and
      6-cylinder  engines and  is:   $22.47 + $35.64 = $58.11
      or approximately $58.
0     This cost  is  $121 per  vehicle  when  applied  only to
      uuGse  ULJ\J ± \ 3  requiring   new   u€CnHGj.ogy  in  mo w£ j.
      year 1988.

The same methodology can be used for LDGT2s:

°,    For  the  3.3  percent of  LDGT2s  which  are 6-cylinder
      engines  and  which  are   converted  from  oxidation
      catalyst   to    three-way    closed-loop    catalyst
      technology,  the  cost   is  $133  per  vehicle,  which
      results in a  contribution of:  0.033 X $133 = $4.43

0     For  the  15.6  percent of  the LDGT2 market  which, are
      8-cylinder  engines  with  oxidation   catalysts  and
      which  will   be   changed   to  three-way  closed-loop
      catalyst  systems  at a  cost  of  $157,  the  resulting
      8-cylinder contribution is:  0.156 X $157 = $24.49


     0     The   8-cylinder   engines   also   have   some  open-loop
           three-way  systerr.s, which  will  go  to closed-loop  in
           model  year 1988.   Of"the  LDGT2 market,  9.5  percent
           will  need new closed-loop control  (8.7 percent  from
           three-way   plus   oxidation  catalysts  to  three-way
           closed-loop   plus   oxidation   catalysts,   and  0.8
           percent   from  three-way  to  three-way  closed-loop
           catalysts)   at   $65  per  vehicle,   which makes   this
           portion  of  the  8-cylinder  contribution:  0.095 X $65
           «  $6.18

     0     The   remainder   of  the  LDGT2  market,  71.6  percent,
      -  -   -already  will have  three-way closed-loop catalyst  or
           three-way    closed-loop    plus   oxidation    catalyst
           systems  in place  by  model  year 1987,  and  thus  will
           incur  no costs.

     0     The  average  hardware cost  per  LDGTZ  in  the fleet
^  	   ^ is the  sum  of  the 6-  and  8-cylinder  contributions
       -~~"~and  is:   $4.43  * $24.49  +  $6.18  - $35.10,  or about

     0     This  cost,  when  distributed only over  those  vehicles
           requiring  new technology,  is about  $123  per LDGT2.

     The  average per  vehicle hardware costs  of  $58  for LDGT(s
 and  $35 for  LDGT2s derived  above  include  the  total  costs for
 hardware   components   added   (three-way  catalysts,    feedback
 carburetor  modifications,   and/or  closed-loop control),  with  a
 credit  for those components  removed  (oxidation  catalysts and/or
 open-loop  control), averaged over the  entire  fleet of  vehicles
 in that  LDGT category.    However,  the  complete cost  of   such
 hardware  should  not  be applied  solely to  the  more  stringent
 NOx   standard,    since   the  manufacturer  also   derives  other
 significant  benefit from its application.  This  is  indicated  by
 the  fact  that  manufacturers  have  already   converted  much  of
 their  fleet   from oxidation to three-way  catalyst systems,  and
 have  stated   in  their comments that  they plan to  continue  this
 trend,  even  though  it  is  not necessary  from  an  emissions
 control standpoint under  the current 2.3  g/mi NOx standard.
 The  application  of this more costly  technology prior  to the
 implementation   of   the  stricter  standard   clearly  indicates
 benefits   to  the  manufacturer,  which  include  improved   fuel
 economy and  driveability   as  well  as  parts consistency  with
 their   light-duty  vehicles.   This  .parts  consistency  leads  to
 greater economic  efficiency  and   lower  total  costs.  It  is
 difficult  to quantify  the   precise value of  these benefits  to
 the   manufacturer;   however,  it   is   clear  that  they  are
 significant.   Absent   any   more  precise  information,  EPA  has
 applied 50  percent of  the costs  to the implementation of the
 standard,  and 50 percent   to  the  other benefits  which  will  be


                       Table 3-8

              Total LPT Manufacturer Cost
RD&T Hardware
-- . $92,350K
$26,970K $476,290K
RD&T Hardware
, — $92,350K
$30,870K $395,950K
Discounted at 10 percent to 1988.

                            Table  3-9
             Light-Duty Trucks  F'irst  Price  Increases

          Vehicles Requiring New Technology

Three-way Closed Loop from
Oxidation Catalyst (LDGT,)
Three-way Closed-Loop From
Oxidation Catalyst (LDGTi » ,)
Three-Way Closed-Loop From
Oxidation Catalyst (LDGT2)
Three-Way Closed-Loop From

$ 68
Three Way Open Loop (LDGTZ)

First Time Application of EGR
Electronically Controlled EGR from EGR
Average for All Vehicles
All LDTs
$ 22
$ 44

$ 31
$ 20
$ 22
$ 44
$ 28


rray experience  a  gain of  up  to  8  percent.   This  was  seen  in
comparisons  of  1985   certification  data  of  matched pairs  of
Federal  and  California  vehicles.   For  fuel  economy  changes
which may  occur,  the  costs  remain  as  in  the Draft  RIA  at $51
for LDGTs  and  $41  for LDDTs lifetime cost  per  affected  vehicle
per  one  percent   change   in  fuel  economy,  either  greater  or
less.   When  apportioned over  all  vehicles,  the cost  is  $21 per
LDGT and $41  per  LDDT, or  $24 per LOT per one percent change in
fuel economy.   It  is  expected  that  on  a fleetwide basis there
may be  a slight gain  in  fuel  economy;  however,  for costing no
fuel gain was included.

     iii.  Total Cost to Users

     To  summarize, purchasers  of  LDTs can  expect  to   pay  an
average  of  $28 more  for  1988 model year  LDTs for  the emission
control  improvements  as compared  to 1987 LDTs.   In the  case of
fuel  economy increases or decreases,  LDGT users  can expect a
$51 change in lifetime operating  cost per  one percent change in
fuel  economy,   while  LDDT  users  can  expect  a   $41  change.   A
slight gain  in  fuel economy is  expected, but  is  not  included in
total cost.

     2.    Aggregate Costs for  the 1988 LPT NOx Standard

     The  aggregate cost  to  the  nation  of complying with the
1988 Federal LOT NOx  emission  regulations consists  of  the sum
of  fixed costs for RD&T  and new emission  control  hardware. No
changes  in  maintenance  or  fuel  economy   costs  are expected.
These  costs  are calculated  based on sales  projections  for the
5-year  period  following  introduction of  the  standard.   These
sales projections were shown  in Table 3-4.

     The  various   costs  associated  with  this rulemaking  action
will  occur in  different periods.   In order  to  make all costs
comparable,  the present  value  at  the  start  of  1988  has been
calculated   based  on  a  discount  rate  of   10  percent.   The
.calculations  were  shown  earlier  in  Table 3-8.  The aggregate
cost  of  complying with  the  new  regulations  for  the   5-year
period  is  estimated  to be equivalent  to  a lump sum investment
of  about  $427  million  (1984  dollars)  made  at the start  of
1988.    When  amortized  over  the  projected   sales  for  the
five-year  period,  the value is  $28  per vehicle  at the  time of


     B.    HDGE NOx Standards

     Comments  on  the costs  of  the proposed  HDGE  NOx standards
were  received  from two  of  the  three  major  manufacturers  of
KDGEs  and  were  not  highly  detailed.   Chrysler   gave  a  cost
estimate  only for  the  later standard,  with  no estimate  for  a
fuel   economy  penalty.   General  Motors   stated  that,   "the
predominant  HDGE  costs associated  with  the more  stringent  NOx
standards  proposed by  EPA  would  be  an  increase  in  fuel
consumption."   Ford comments did  not  discuss  the  costs  of  the
standard.  Therefore,  any cost  revisions  below are  based  on a
reanalysis  of  the  control  technology  necessary,  which  the
manufacturers'  comments  did discuss,  rather .than  on concerns
for  cost  estimates  for   specific  components  of  the  control

     Before  beginning   this  reanalysis  of  the costs  to  comply
with the  1988 and  1991 HDGE NOx standards,  a  brief discussion
of - HDGE   certification   data   and   options  and   potential
certification  approaches  is  necessary.   First, as  of February
1985,  the three major  HDGE  manufacturers  had certified a total
of seven  families.   This is  a  decrease of  eight  families from
1984,  brought  about   by  IH leaving  the  market  completely,
Chrysler  dropping  one  family,  and GM  and Ford  each combining
two  families which were  previously separate.   However,  due to
the split class HDGE emission standards beginning  in 1987,  the
number of HDGE families  is  projected to  increase from 7 to J.O
or  11  even though  no  new engine  offerings  are  expected.   For
simplicity,  and since  all HDGEs will have  to meet the same NOx
standards, this analysis will assume  that  HDGE sales are spread
evenly  among the  11 families.   This  allows fixed  costs  to be
assigned  on  a per  family  basis  and  spread  over  the  entire
fleet,  without  having  to  assign  specific  fixed  costs  to
specific  families  for  the  sake  of  production-weighting  the
fixed  cost  impacts.   As  will be  seen  later,  in  the long term
this introduces no  error  into the per vehicle cost.

     Second,   it  is  worth   noting  that  beginning  in  1987,
manufacturers  may  exercise the option to certify  their HDGVs of
up to  10,000  Ib GVW as  LDTs.  While this  option also exists for
the  1988  model  year,   this analysis  does  not  evaluate  that
possibility.   Presumably  it  would  be more  expensive on  a  per
engine basis  to meet the LOT NOx standard  (probably  requiring a
three-way catalyst  with  closed-loop  control)  than it  would be
to  meet   the   1988 or  1991 HDGE standards.  -Therefore,  if
manufacturers  choose to exercise this option in  1988,  it would
be  based  on  their belief  that  other perceived  or  intangible
benefits  are  worth  any  extra costs.

     Third,  any analysis  of  costs to meet  the 1988 or 1991 HDGE
NOx  standards must  be  placed in  the proper context by  reviewing
current  HDGE NOx  certification  levels.   The certification data


presented  in   the   HDGE  technological   feasibility  analysis
indicates chat none of  the  families certified in 1985  .T.eet the
1988  6.0 g/BHP-hr NOx  standard,  even though two configurations
within  these  families  do  meet the 1988  standards  anc  one  of
these cwo configurations meets  the 1991 standard.  For purposes
of cost  estimation,  this analysis  will assume  that  no current
HDGE  families  meet   the  1988  NOx  standard  of 6.0  g/BHP-hr.
However, several will  be very close.

     1.    1988 NOx

     a.    Cost to HDGE Manufacturers

   •  i.    Fixed Costs

     As  noted  above,  no  comments  specifically addressed  the
issue  of  manufacturers'  cost of  the  intermediate   HDGE  NOx
standard. •-However,   a  reanalysis  of  cost  has been  done  to
reflect changes  in the  EPA projection of the control technology
necessary  to  meet  the  standard,   as  discussed  in  Chapter  2,
Technological   Feasibility.    These  new   cost   estimates  are
outlined below.

     The   costs   originally   estimated   for    RD&T   included
recalibration of  the  fuel,  ignition, and EGR systems as well as
certification  costs.   In  total,  these  recalibrations  amounted
to  $39,600   per   engine   family   based  on  three  calibration
combinations at six person-weeks  of effort each.  Certification
costs  in  the  Draft RIA were $192,170  per  engine family, based
on one durability and three data engines per family.

     Ths  final  cost  estimate  includes  these costs,  plus costs
for  evaluation  and   recalibration  for  improved secondary air
management,   redesign   of    the   combustion   chamber,   and
emission-related   improvements   of   the   intake    manifold.
Secondary  air  management recalibration  is expected  to  require
about  the same  level  of effort  as the fuel, EGR,  or ignition
system, or $13,200 per  engine  family.  Redesign  and  testing of
the combustion  chamber is expected  to  consist  of a redesign of
the cylinder  head and  was  estimated in  the  Draft  RIA for the
proposed  4.0  standard  to  cost   $306,900  per   engine  family,
(including  a  10  percent  contingency  factor) .   This  value  is
used  here.   Finally,  enhancement  of  the  intake  manifold  is
estimated  to  require  about  five  times the  level of  effort  of
any. of  the  above recalibrations,  an amount  of  $66,000 per
engine  family.

     As  was  discussed  above,  it  is  now  projected  that the
major manufacturers will  certify  a  total of 11  HDGE  families in
1988.   The  three tasks  originally  presented  in the  Draft RIA
are expected  to be necessary for  all 11 engine  families.   Total
cost  is  thus  $580,800  for  recalibration  of  EGR,   fuel,  and
ignition systems.


     Of  the 11 engine  families receiving this recalibration,  8
are  expected  to   be able  to  rr.eec  the  6.0  standard  without
additional  changes.   This  is  based on  a  review  of current  NOx
certification  levels which indicates  only  three  families  have
NOx  emission levels  above  8.0 g/BHP-hr.  The  other three  HDGE
families  may require some or  all of  the additional work  listed
above:    secondary   air   recalibration,   combustion  chamber
redesign,  and  intake manifold  improvement.   As  given  in  the
Draft  RIA,  RD&T  costs  for  applying  these  improvements  to  each
of  the remaining  engine families are  $13,200  for  the secondary
air,  $306,900 for the  combustion chamber,  and $66,000  for  the
intake manifold  for  a  total  of  $386,100  per  family.  Assuming
that   all   three   families  do all  the  work,, this  totals  an
additional  $1,158,300.

     Certification  must  be   conducted  for  all   11    engine
families.   Using the  $192,170  certification  cost  per   family
presented   in  the   Draft  RIA,   certification   costs    total
.$2,113,870.   The  total  fixed  costs  to meet  the  1988 HDGE  NOx
standard  is the  sum  of  the development and certification  costs
or  about $3,708,000.  The separate  components of  these  costs
are  detailed in Table 3-10.

      ii.    Variable  Cost

     EGR  is the  major  NOx emission  control component expected
on  HDGEs.   A review  of  the 1985  certification  records
that   six   of   the  seven  HDGE  families   currently  have  EGR
installed.   One  family,  representing about 4  percent of  sales,
would  have  to install EGR  to  meet the  1988 HDGE NOx standard.
In  a  1980  study  completed for EPA,  HDGE  EGR  was  estimated  to
cost   $9.36  at   the vendor   level  (1977   dollars).[10]    When
adjusted  for  inflation  to 1985  dollars using the new  car  CPI
 (1.43) and  accounting for manufacturer  and  dealer  overhead  and
profit (1.29), [9] HDGE  EGR is estimated to have  a  retail  price
equivalent  of  $17.27. When spread  over all the vehicles  in the
fleet, this averages $0.69  per vehicle.

      In addition the recalibration  discussion above indicated
that 3 families may  need additional work for combustion  chamber
modifications   and  intake  manifold  'improvements.    Redesigned
hardware  may be  necessary  for the  three HDGE families  needing
this  work.   In   the long  term,  the  redesigned  parts  could
presumably   cost   the  same  as  those  being  replaced,   but  a
conservative  approach   is   taken   and  $10   is   assigned  per
 redesigned  engine.   When this  cost  for  work on   three  engine
 families   is  spread  over  all  HDGEs,  the  hardware cost  per
 redesigned  engine due to  the   &.0  standard  is  $2.73.  When the
EGR cost is  added  to the  redesigned component cost, the total
hardware  cost sums to $3.42 per HDGE.


                      Table 3-10

         Summafy of  1988  HDGV  NO*  RD&T  Costs

                             Families    Cost per
	Category	     Affected     Family     Total

Fuel, ingition and EGR        11        $39,600   $435,600

Secondary air management,       3       $386,100 $1,153,300
combustion chamber redesign,
and intake manifold mods.

Certification                 11       $192,170 $2,113,870



     iii .   Total Manufacturer Cost

     The  total  manufacturer cost of  the 6.0 HDGE  NOx standard
is  the sum  of  the  RD&T  cost  and  the  hardware  cost  for  the
engines produced in  the  three  model  years immediately following
introduction of  the standard,  all  discounted at 10  percent  to
1988.  Projected sales have  been  updated to  reflect information
in Reference  3,  and  are presented  in Table  3-11.   These sales
figures  have  been  used  to generate  the  total  manufacturer
hardware  costs,   and  are   presented  together  with total  RD&T
costs  in   Table   3-12.   Manufacturer  costs  are  shown  to  be
$7,671,000 undiscounted  and  $7,869,000 discounted.

     b.    Cost to Users

     i.    First Cost

     Manufacturers  must  recover  their  costs by  increasing  the
first-price  of vehicles equipped with  HDGEs.   It  is  expected
that  manufacturers  face  a  10  percent  cost  of  capital  and
recover their  RD&T costs  in the three  model years immediately
following   introduction   of   the   standard,   1988-90.    The
discounted RD&T  costs amortized  over  the engines  projected  to
be sold in those  three model years results in a cost per engine
of $4.02,  or about  $4.   The sum of  the  engine  share  of  RD&T
cost   and  the  hardware  cost   ($3.42)   is  the  first  price
increase.   Averaged  over  all  model year 1988-90  HDGES,  the
total is $7.44, or about $7  per engine.

     ii.   Operating Costs

     As described  in  the  technological  feasibility  analysis,
the  fuel  economy  impact  of  the 6.0 NOx  standard  is  expected  to
be negligible  for HDGEs.   Since  only  engine  recalibrations and
component  redesigns  will   be  used  to  achieve  the  required
emission  reductions,  maintenance  should  not  be  affected by this

     iii.   Total User Cost

     The  total cost  to  the user  is  simply  the  first  price
increase  of  approximately  $7  per  vehicle  equipped  with  an
HDGE.  Operating'costs are not expected to change.

     2.    Total  Manufacturer  and.  User  Costs ' for  the  1988

     a.    Manufacturer  Cost

     The  total manufacturer cost  of  compliance  for  the  1988
HDGE  NOx  standard of  6.0   g/BHP-hr  for  the  three  model  years


- - 1994
Pro1 acted
Table 3-11
[ HDE Sales (thousands)

Based on information presented in Reference 3.


                           Table 3-12

         HDGE Manufacturer Costs for 1988 NOx Standard
RD&T Hardware



        $3,708K       $3,963K

Total           $7,671K                          7,869K
*    10 percent to 1988.
**   Research and development costs
***  Certification costs.


 1988-90  is  the surr of fixed and variable costs developed  above,
 and   is   about  57.7 lion  undisccunted   or   $7.9  million
 discounted  at  10 percent  to the year of the standard.

      b.     User Cost

      The  user  cost  is   the  sum  of  the  first  price  increase
 developed  above  and  any  change  in operating  costs due  to  the
 standard.   No operating  cost increases  are  expected,  so  that
 the  average cost  to  the  user of  a model year  1988-90 vehicle
"with  a HDGE is about  $7.

      3.     1991 NOx

      a.     Cost to Manufacturer

      EPA  received  only one comment concerning  the  cost  estimate
 per   HDGE   due  to  the proposed   4.0  NOx  standard.    Chrysler
 estimated   a   cost   of  $180  for   reduction   from   6.0  to   4.1
 g/BHP-hr,   compared   to   the  estimate   in  the  NPRM   of   $18.
 However,  Chrysler's  comment did not detail the technology which
 would cause this price increase, nor  did it indicate  the  amount
 of  research,  overhead and markup  contained in the  estimate.   It
 is   therefore  difficult   to  determine   to   what   extent   the
 difference   is  based on  actual  differences  between  EPA   and
 Chrysler  estimates of specific  costs, and to  what  extent  it  is
 due    to    differences   in   assumptions  in   areas   such   as
 technological  approach, mark-up,  vehicles over  which costs  are
 apportioned,  etc.

      Nevertheless,   absent   any   detailed  comment,   EPA   has
 ra_evaluate<3  the cost of  the 1991 HDGE  NOx  standard based  on
 the   revision  of  the  HDGE  NOx   portion  of  the  technological
 feasibility analysis.

      i.     Fixed Cost

      The  cost analysis for  the proposed  4.0  HDGE NOx  standard
 contained   in  the  Draft  RIA  included  RD&T  costs  for   the
 recalibration  of   the   fuel,   EGR,   and   ignition   systems,
 combustion  chamber  modifications,  arid  for  certification.   The
 final cost  analysis for   the  1991 HDGE  NOx  standard  includes
 RD&T  costs  in   these  areas  plus  others  for  improvement  of
 secondary  air management  and intake  manifold modifications  for
 those HDGE  families not already receiving -these  changes.  These
 costs are allocated  as described below.   The per family cost  to
 accomplish  each of  these  tasks  is  the  same  as allocated  for
 1988.  For  convenience,  these  costs are  shown  again  in Table


                           Table 3-13

                   RDS.T Costs per HDGE Family

                                                    Costs per
     	Tasks	                        Family

1.   Fuel, ignition and EGR recalibration             $39,600

2.   Secondary air management                         $13,200

3.   Combustion chamber redesign                     $306,900

4.   Intake manifold modifications                    $66,000

5.   Certification                                   $192,170


     First, coses  are  again allocated  to  each HDGE  family for
further fuel,  ignition,  and EGR  recalibration work.   However,
this is  probably conservative since it  is  reasonable to expect
that some  families will  be able  to  meet the  1991  NOx standard
with only minor changes to 1988-90 calibrations.

     Second,  further  costs  for  the  more   significant  changes
(secondary  air  management,  combustion  chamber  redesign,  and
intake  manifold  modifications)   are  now   allocated   for  these
eight families not receiving  these changes  in  meeting the 1988
standard.  This  is also  conservative,  since it is unlikely that
all  eight  families  would  require   all   three   of   the  more
significant  changes.    Thus,   as   is   shown  • in  Table  3-14,
recalibration  work  for all 11  families  totals to  $435,600 and
other more  significant changes  for the  remaining eight families
totals to $3,088,800.

   -  Third, certification  costs are  once again appropriated for
all  11  families  at  a total  cost  of  $2,113,870.  Once again,
this  is  conservative, since  it  is  likely  that  some families
will  be  able   to gain   1991  certification  through   a  running
change in  lieu of full certification.

     As  shown   in Table  3-14  when  the  work  is  allocated  as
discussed   above  and  summed,   RO&T   costs  total   $5,638,000

     ii.   Variable Cost

     As  was   discussed   above,    redesigned   hardware  may  be
necessary  for  the combustion  chamber  and  the  intake manifold
modifications,   and  was conservatively  estimated  above  to cost
$10  per  affected  engine.  For  the 5.0  standard,  8  of the 11
HDGE families  will  require this  hardware.   Spreading this cost
evenly over all  HDGEs, the average cost  per engine  is  $7.27.

     iii.  Total Manufacturer Cost

     The  total manufacturer cost,  including RD&T and hardware,
sums   to   $13,933,000   undiscounted    and   $14,153,000   when
discounted  at  10 percent to 1991.  This includes the  RD&T  costs
developed  above  and  the  costs for redesigned  hardware on  model
year  1991-93   engines.   The  stream of  costs,  both undiscounted
and discounted,  is shown  in Table 3-15.

     b.    Costs  to Users

     i.    First  Cost

     The incremental increase in  the first  price  of a 1991 HDGV
over  a'similar  1990  HDGV can best  be presented  as the average


                      Table 3-14

          Sumrr.acv of  1591  HDGE  NOx 3D&T Costs
Fuel, ingition and EGR
Secondary air management,
Families Cost per
Affected Family To
$386, 100
combustion chamber redesign,
and intake manifold mods.

Certification           " "  11       $192,170 $2.113,870



                          Table 3-15

         HDGE Manufacturer Costs for 1991 NOx Standard
2*7 "Jr\ V
, 1 /UK
RD&T Hardware
$6,589K $7,564K




1993 -
Total          $13,933K                        $14,153K
*10 percent to 1991.
**   Research and development costs,
***  Certification costs.


first  price  increase  expected  if  costs  are  spread  ever  all
HDGEs.  If  RD&T costs  are  amortized  over three  years  of  sales
(1991-93)   at  a  10  percent  cost of  capital,  the  average  per
engine  increase  attributable to  RD&T  equals  $6.33.   This  added
to a  fleet average hardware  cost of  $7.27  gives a short  term
average  first  price  increase  of  $13.60  for  the  1991  NOx
standard.   In the long term this cost drops to about $7.

     ii .   Operating Costs

     As is discussed  in  the technological feasibility analysis,
no significant  fuel economy impact  is expected for HDGEs due to
the 5.0 standard.   Therefore,  fuel  costs will  not  be  affected.
Increased  maintenance is  not  expected  as  a  result  of meeting
the  1991  HDGE  NOx  standard,  so maintenance  costs should  not

     iii.   Total User Cost

     The  total  user  cost of  the 1991  standard  is  simply  the
first cost  increase,  averaging $13.60 over the cost of  a  model
year  -1990  vehicle  equipped  with  an  HDGE.   No  increases  in
operating costs are expected.

     4.    Total  Manufacturer   and   User Cost  for  the  1991

     a.    Manufacturer Cost

     The  total manufacturer cost  of compliance for  the 1991
HDGE  NOx  standard  of 5.0  g/BHP-hr   for  the  three  model  years
1991-93 is  the  sum  of fixed and variable costs developed above,
and  is  about  $13.6  million   undiscounted   or  $14.2  million
discounted at 10 percent  to  the  year of the standard.

     b.    User Cost

     The  user  cost  is   the  sum of   the  first  price  increase
developed  above and  any  change  in operating  costs due  to  the
standard.   No operating  cost increases  are  expected,  so that
the  average  cost  increase to  the user of a  model  year 1991-93
vehicle with  a  HDGE is $13.60.  After RD&T costs are amortized,
the  first price increase  will  drop to  about $7  per HDGV.

     C.    HOPE NOx and Particulate Standards

      Specific  comments  on the proposed HDDE  NOx  standards were
received   from  five  manufacturers  —  Cummins,  Ford,  General
Motors, International Harvester, and Mack, as  well  as  from the
Department  of  Energy, the Engine Manufacturers Association, and
the  American Trucking Association.   All commented  either that


hardware costs were  substantially higher or fuel economy losses
considerably greater, or both, than  those  estimated by EPA.  In
general, detailed  derivations of  the  costs  were  not  given in
the   comments,   and   EPA's   derivation   approach   was   not
challenged.   However,  in one case a  comparison  of assumptions
was  made  which  detailed  the reasons for  fuel  economy   cost
estimate differences without  investigating the  relative  merit
of the two methods.

     The   comments   have    prompted    revised   analyses   of
manufacturer cost  estimates  by EPA.    In the case of RD&T, costs
are  based  on  the number of  engine families which  will require
the  work;  at  this  time, family-specific  data- on  HDDEs  is not
available.    Therefore,   EPA   can  only  estimate  the  number  of
families which will  require  RD&T allocations  based  on general
manufacturer  comments.   For  hardware  costs,   those  components
which were not costed by Rath & Strong[1] were estimated by EPA
in  the  Draft  RIA;  again,   they  are  updated here    based  on
general  manufacturer  comments.   These  estimates  are  retail
price  equivalent  (RPE)  costs.   The  detailed  reanalysis  is
provided in the following sections.

     1.    1988 NOx  Standard

     a.    Cost to HDDE Manufacturers

     Hardware  changes  deemed necessary  for engines to meet the
6.0  g/BHP-hr NOx standard as  outlined  in the Draft RIA included
injection  timing  retard and the  addition of  aftercooling to
non-aftercooled  turbocharged engines.   In comparison  to this,
Ford  outlined  hardware  plans   of   improved  fuel   injection
systems, variable injection  timing,   and  turbocharging  on all
engines, as well  as  charge air cooling on  some.  Cummins listed
variable   injection   timing,   low   temperature    aftercooling,
increased    fuel    injection   pressure,   combustion   chamber
modifications,  and  an  electronically controlled  fuel system,
while  International  Harvester   listed  engine  cooling  system
•changes, air-to-air  aftercooling,  and electronically controlled
fuel systems.   The  dissimilarities  in  these  lists of hardware
changes contributed  to  the  difference  in cost estimates between
manufactures   and  EPA;  they   also   prompted  a  revision  of
development  tasks  and hardware  in the  EPA analysis,  which is
presented  later.

     The costs presented by manufacturers  for  HDDE 6.0 g/BHP-hr
NOx  control are compared to EPA's  projections  as  follows:

        Retail  Price  Increase  per  HOPE,  6.0  NOx Standard
Draft Analysis     $16


Cummins       $100-800
           HDDE average

           HDDE wich new technology
           on   need
        for   variable
HDDE average,

HDDE average
including particulate
-•	Ford's  cost  is  an  estimate of the  consumer  cost for  the
 hardware changes described  above,  which  are  currently planned
 for   model  year  1987  in  anticipation  of  more  stringent  NOx
 standards  for that year.  Cummins  indicates  " increase  in
 estimated  engine  prices  for the  Cummins product  line."   The
 value given  by International Harvester (IH) is  a sales-weighted
 value of going from  10.7  to 6.0  g/BHP-hr NOx based on the  cost
 difference  between present  California  and Federal  IH engines.
 The  California NOx standard  is  5.1  g/BHP-hr.

      The costs  presented by manufacturers  are clearly  higher
 than  those  given  by  EPA,  prompting  the reanalysis  which  is
 included below;  however,  since  the industry  estimates  do  'not
 give  detailed  breakdowns  of   components  and  costs,   it   is
 difficult  to  tell whether  the  values  in the  comments  can  be
 directly compared to  the  EPA  estimates.    For   example,   the
 industry estimates  are   presumably  per  engine  requiring  new
 technology,   although  this   is   not  clearly  stated;  the  EPA
 estimates  presented  in the  NPRM  are  spread over all  HDDEs,  some
 of which are  projected not  to  require  the new  hardware:   Also,
 manufacturers, when  indicating  increase  in  "consumer  cost",  do
 not  indicate which RD&T  costs  are  included,  nor  the  amount  of
 dealer markup.   From aggregate  RD&T  costs  which were provided
 confidentially by  some manufacturers,  it  appears  that ongoing
 basic research costs are  included,  rather than only  those costs
 which arise  from research directly applicable  to this  standard,
 as in the EPA analysis.

      Dealer   markup   is   presumably  higher   in   the   industry
 analysis than  in  EPA's analysis;  EPA  bases   its markup  on  the
 idea  that  the dealer  will  incur no costs due to the standard
 except for  the interest that must be paid on  the higher cost of
 the   inventory  before   it   is   liquidated;   this   interest  is
 included in   the  EPA  markup value.   Using  this  method,  the
 dealer will receive  no  profit  due  to the standard,  but will not
 take  a  loss  either.   On  the other hand,  if  the  manufacturers


are  including  their  usual  dealer  markup  in their  "consumer
cost"  estimates,  the  dealer  is  caking  a  profit  from  the
standard;  such  dealer  profit  is  not  correctly applied  to  the
cost c.f the standard.

     Such differences  in the analyses  may partially explain the
differences  between  EPA  and manufacturer  cost estimates,  and
create  a  situation  where   the  values  cannot   be  directly
compared.    Resolution   of   these  potential  differences   is
confounded by the fact that the development  of   the  industry
cost estimates  is not documented  in the  comments,  so that even
the discussion  above  is only a  conjecture  as to what  the cost
values presented in the comments may actually represent.

     Nevertheless, as  part  of  its review of  the  technological
feasibility   of   the   6.0   g/BHP-hr   NOx  standard,   EPA  has
reevaluated   the  control   technology   needed   to  meet  the
standard.  This in turn has  led  to  a  reanalysis  and revision of
the cost figures; this revision is discussed below.

     i.    Fixed Cost

     The  reanalysis  of  the  RD&T  and  hardware costs necessary
for HDDEs to comply with the 6.0  g/BHP-hr NOx standard includes
the  timing  retard and addition  of  aftercooling as  in the Draft
RIA,  plus  additional  RO&T  and  hardware  costs  of  improved
aftercooling,    variable    injection    timing,   and   improved
turbocharging.   The number  of  HDOE families remains at  86,  as
in the Draft RIA.

     Timing  retard calibration  evaluation was costed at $26,400
per engine family in the Draft  RIA, based on three calibrations-
per engine family, 160  labor hours per calibration at a rate of"
$50 per  hour,  and a  10 percent  contingency  factor.   This value
has been  increased to  $132,000, five-fold the original, because
an   increased   number   of   calibration   evaluations   would  be
necessary  to optimize fuel  economy and  to  deal  with the  larger
number  of approaches  available  for   meeting  the  6.0  g/BHP-hr
standard.  For  1986  engine  families,  RD&T comes  to  $11,352,000
for  timing retard.  The addition of aftercooling  to  10 percent
of  the  HDDE  families  (half  of  those  turbocharged  engines
without  aftercooling)  remains  as before, at  $57,400  per  engine
family  and  a  total  of  $494,000 based on  six person-months of
engineering  and development  work  per family.

     New   to   this  analysis  for  the   6.0   standard   is  the
improvement  of  aftercooling, which was  previously believed to
be  necessary only for  the  4.0  standard  but is  now  added  in
response  to  manufacturer  comments.  This RD&T cost also remains
the  same as   in  the  4.0 portion  of the  Draft  RIA,  at $172,200
per   family   for  air-to-air   and  $57,400  for   air-to-liquid


aftercooler systems.  At a 50 percent  application  rate  for each
system   for   the   72   engine_  families   expected   to   have
aftercooling, the total cost is $8,266,000.

     Variable injection timing  (VIT)  and  improved turbocharging
are also  new to  this  analysis,  and  are  each  estimated  to have
RD&T costs of $95,700 per engine  family;  however,  half  of these
costs are  applied to the particulate standard,  leaving  $47,850
per family  for  each of  the  two tasks.  This value is based on
two designs  per change, four person-months  per  design,  and the
usual $50 per  hour and 10  percent  contingency  factor,  as well
as an additional  25 percent  to account for the effort needed to
optimize fuel economy.  Assuming  VIT  and  improved turbocharging
will each be assigned to 50 percent  of the  86 engine families,
these costs are $2,058,000 per tasK, or $4,115,000.

     Certification  costs  remain at $6,500,000 as presented in
the  Draft  RIA.   This  includes  dynamometer  time and  emission
test costs for one durability and three data engines  per engine

     Total  RD&T costs are then  the sum  of  all  these costs, or
$30,738,000.   This  is  comprised  of  $11,352,000  for  timing
calibration   evaluations,    $494,000    for  the   addition   of
aftercooling,   $8,266,000   for   the  upgrading   of   current
aftercooling  systems,  $4,115,000  for  VlT/improved turbocharging
and $6,500,000  for certification.

     ii.   Variable Cost

     Hardware  costs  per engine  have also   increased.   While
costs for  injection timing  retard  and addition of aftercooling
remain,  total hardware  costs are increased due  to  the  addition
of.   improved   aftercooling,   variable   timing,   and  improved
turbocharging.   The per engine hardware  cost  for HDDEs  adding
aftercooling  capability  remains  at   $61,  with  10   percent of
KDDEs being  affected.

     Improved  aftercooling  cost  per  engine  also  remains  as
originally  in the  Draft RIA for the  proposed 4.0 standard, at
$73  for  conversion  of  an  air-to-liquid   to  an  air-to-air
aftercooler,  and  $91  for  upgrading of an existing air-to-liquid
system.  These  costs,  however,  are now being  allocated  to  some
engines which will be built under  the 6.0 standard,  in response
to manufacturer comments that some  will need the technology for
the  earlier  standard.   The  rate  of  application is  such  that
half of all  turbocharged engines  (31  percent of  all  HDDEs)  will
employ  new  or  upgraded   aftercooler  systems  for  the   1988
standard.   One-third  of  this 31  percent,  or  10  percent  of the
total,    is   comprised   of    HDDEs   getting   air-to-liquid


aftercooling for  the first  time as  described  above.   Of  the
remaining  21  percent,   16   percent  will  convert  to  improved
air-to-liquid  aftercooling   and  5  percent   will   convert  to
air-to-air  aftercooling.   These costs  average to $87  for  each
vehicle  converting  to   improved  aftercooling,   and  $18  when
applied to all  HDDEs.

     The  incremental cost of electronically-controlled variable
injection timing is estimated at $25 per engine,  and is applied
to half  of  the engines, with half  of  the  cost charged  to the
particulate standard.  As noted  above,  this  is  an EPA estimate
based on manufacturer comments to the NPRM.

     Improved turbocharging  is  estimated to cost  $5 per engine
as in  the  Draft  RIA,   and  would apply to  50  percent of  all
turbocharged HDDEs  (31  percent  of  all HDDEs).   Half  of  this
cost  would be applied to the particulate standard.

     The  sum of  these  costs on a   fleetwide  average  basis  is
about $32 per HDDE.  When  applied only to engines requiring the
new  technology,   and  the  average  hardware  cost  for  the  6.0
g/BHP-hr NOx standard would be $93 per engine.

     iii.  Total Manufacturer Cost

     Total  RD&T  and  hardware  costs  must be  discounted  at  10
percent  to  the  year  of  the  standard,   1988,  in  order  to
represent actual  manufacturer cost.   It is  reasonable to expect
that   RD&T   expenditures  will   be   made   in   the   two  years
immediately  preceding   the  year of the  new  standard.   The RD&T
costs described  above  are  summed in Table 3-16,  and presented
in  undiscounted   and   discounted  form.    Hardware   costs  are
allocated  according  to  sales  projections,   which  have  been
updated  due to  new information[3]  and  were  shown  in  Table
3-11.   Using these sales projections  and  the average  cost per
HDDE developed  above results in the  distribution  of hardware
costs shown  in  Table 3-17,  in both undiscounted and discounted
form.   Since the 6.0  g/BHP-hr  NOx  would  only  apply through
1990,  the  hardware  costs  are   presented  and  summed  for  only
three model years of HDDE  sales.  Manufacturer hardware costs
sum  to  $34  million   dollars   undiscounted   and  $31  million

     These   manufacturer   costs  for   RD&T   and  hardware  are
summarized  and presented  on  an annual  basis  in  Table  3-18.
This analysis results in a  total undiscounted manufacturer cost
of about $64.4 million and  a  discounted cost  of  about  $66.4


                                Table 3-16

                        HDDE  RDS.T  Co'sts  got  6.0  NQx

        	Undiscounted	  	Discounted*	
            Non-                              Non-
        Cert. Costs  Cert. Costs  Total  Cert. Costs  Cert. Costs  Total

1986     $17,OOOK    .Sl.OOOK   S18.000K   $20,570K     $1,210K  S21.780K

1987     $ 7,200K    $5,50QK   SI2.7QOK    $7,920K     $6,050K  $13,97QK

TOTALS  -$24,200K    $6,500K   $30.700K   $28,490K     $7,260K  $35,750K
     Discounted at 10 percent to 1988.


                      Table  3-17

      HDDE Hardware Costs for 1983 NQx Standard

. S30.609K
Discounted at 10 percent to 1988.


                      Table 3-18

     HDDE  Manufacturer Costs  for  1988  NOx  Standard
	 TOTAL . .
Und is conn ted
RD&T Hardware
$30.700K $33,656K
Di scounted*
Hardwa re
" ' 10,207K
Discounted at 10 percent to 1988.


     b.     Cost to Users

     i.     First Cost

     Increases  in  HDDE  purchase  price  due   to  the  6.0  NOx
standard  are  determined  in  the  same  manner  as  for  the  LDT
standard, except  that  the capital costs  (RD&T)  are  expected to
be recovered  in  three  rather  than five  model  years due  to the
introduction  of  the  second NOx standard in 1991.   The average
increase in first cost of HDDEs would consist of the sum  of the
discounted  RD&T cost  amortized  over  vehicle  sales  for  model
years 1988  through  1990  through  plus  the  average  per  engine
hardware  cost.   These  costs  can also  be  expressed   per  HDDE
requiring new technology rather than as  average per  engine cost
by  adding  the  RD&T  cost  apportioned  only  over  the  affected
vehicles  to the cost  of  the  hardware  required.    Costs  using
these two different approaches are presented below.

     First, total discounted RD&T cost  amortized over the total
HDDE sales  projected  for  model years 1988 through  1990 results
in a cost  of  approximately $37 per  vehicle.   When this is added
to the  $32 average hardware  cost developed above,  the average
first price increase is $69.  When  distributed only over those
engines  affected  by  the standard,  the  costs  are  $50  for  RD&T
and  $93  for   hardware,  for   a  total  of  $143 for  a vehicle
receiving new technology.

     ii.   Fuel Economy

     In  Chapter  2, Technological  Feasibility,  it  was  estimated
that fuel  economy  penalties  associated  with the 1988 model year
NOx  standard  would be  in  the tange c£  C to  2 percent  in the
short term.   This  penalty should tend to  disappear  by the time
of  implementation  of the  second  standard, 1991,  as part  of  a
normal  trend   toward  further  engine and  vehicle  improvements.
EPA  has  reevaluated  the  cost  impact of these  short-term fuel
economy   losses  based   on  the  comments  received.   This  is
presented below.

     First, fuel economy  estimates for 1988  HDDVs  have  been
updated,  and  are  derived  from  information   in  Reference  3.
These  estimates  for  LHDDEs,  MHDDEs,  and  HHDDEs  have  been
lowered  to 15.1 mpg,  8.0  mpg, and  5.9  mpg  respectively.  This
makes them  closer to  the  Argonne  National Laboratory estimates
supplied in  comments  received  from the  Department  of  Energy
(DOE)  which  compared  EPA  and   ANL  assumptions.   These  fuel
economy  values  are   combined  with  a  fuel  cost  of  $1.20 per
gallon;  ANL used $1.46 as  the estimated average  cost  over the
lifetime of  the  vehicle.   However, since the  price  of  diesel
fuel  has varied  significantly in  the   recent  past,  and   since
fuel  prices  continue  to  be  highly   sensitive  to  the  world


political climate  and,  therefore,  unpredictable,  EPA  has  used
S1.20/gal as  representative  of ^today's  price without attempting
to project future price increases.

     Average  annual  mileages  and  lifetimes  remain  as  in  the
Draft RIA,  at 11,000 miles per year for  10  years,  30,000 miles
per year for  9 years, and 65,000 miles per year  for  8 years  for
LHDDEs,  MHDDEs,  and  HHDDEs,  respectively.   These values include
one rebuild  for  some of the MHDDEs and most  of  the  HHDDEs,  and
are  reasonable  estimates  of  the  actual   lifetimes   of  these
engines  for  fuel - economy  purposes.   The  useful  life  VMTs  used
by ANL do not involve any rebuilds, but EPA  has found that  the
majority of  the  heavier HDDEs are not,  in  fact,  retired after
their initial useful life,  and  hence would  continue  to accrue
fuel  economy  penalties.    Therefore,  EPA  has  included  these
higher lifetime  values  in calculating the lifetime fuel economy
cost for the  standard.

     A  10  percent  discount rate  is  employed  with   the values
given above  and the  fuel  economy estimates  are sales weighted
in  arriving  at  the  average  cost  per  engine and  total average
lifetime cost.  The sales  fractions  used  are 35 percent LHDDE,
29  percent  MHDDE,  and  36  percent HHDDE,   and are derived  from
information presented in Reference 3.

     Using  the  fuel economy,  fuel  price,   and vehicle/engine
average lifetime miles  and years  it can be calculated  that e.ach
one percent  reduction  in  fuel economy  corresponds  to  an annual
increase in  diesel  fuel usage of 7.3 gallons  for  LHDDEs,  37.5
gallons  for  MHDDEs,   and  110.2  gallons  for  HHDDEs.   These
increases in  fuel usage correspond to lifetime  discounted costs
Of  $54  for  LHDDEs,  $259 for MHDDEs, and $705 for HHDDEs.  Sales
weighting these  costs  gives the  average  lifetime cost for  a  1
percent change in fuel  economy of  $348 per affected engine.

     Applying these  average  costs,  the   range  in   the  fuel
economy cost  per engine which corresponds  to  the 0  to  2 percent
change  expected  for model year  1988  vehicles for the  1988 HDDE
NOx  standard  is  $0  to $696."  This  value   should  drop to  $0
before implementation of the 1991  standard.

     iii.  Maintenance

     No  increase or  decrease in  maintenance  is  expected  as  a
result of the application of the  technology  needed  to meet  the
6.0 'g/BHP-hr NOx standard  and hence there  should  be  no impact
on  costs.


     iv.    Total  User  Costs

     In summary,  owners  of  model year  1988 through  1990  HDDVs
can be expected to pay  an  average  of approximately $69 more for
the emission  control  on  these vehicles  than  they  would  have
paid without  promulgation  of the  NOx  standard.   In  terms  of
fuel costs, the increased average  lifetime  cost per  vehicle  is
expected to be between  $0 and $696,  tapering off to $0 in later
model years.  Total  lifetime increase  is  thus  $69  to $765  in
the short term,  and $69  in the long term.

     2.    1988  Particulate  Standard

     a.    Cost  to  HOPE  Manufacturers

     i.    Fixed  Cost

     The RD&T costs  for the  non-trap particulate  standard were
reevaluated, but due  to  the lack of specific comments, EPA saw
no  need  for  major change  from  those  costs  presented  in the
NPRM.  Some  revisions in the 1988  particulate RD&T  costs are
caused by  changes  in  the RD&T costs for  1988  NOx  control which
are  allocated equally  with  particulate  control,  and  general
comments indicating the  need  for more development to deter fuel
economy penalties.

     The original non-certification RD&T cost was  based on four
tasks:   1)  modifications  to the   combustion   chamber  through
changes  in the  piston,   2)  changes  in  injectors  and increased
injection  pressure, 3)  changes in  the  fuel delivery system to
refine air/fuel  ratio control during  transient operation,  and
4)  changes   to   the   turbocharger   to  improve  air  delivery
characteristics during  transient operation.  In the  Draft RIA,
the cost per  task  to  accomplish  this non certification RD&T was
estimated   at  $3,292,000.     This   was  based  on   2  design
evaluations  per  task,  4 person-months per  design   at $50 per
hour, and  a  10 percent  contingency factor,   applied  to one-half
of  the 86  engine families.   EPA  determined that one-half of the
families would need the work based on  manufacturer  comments to
the NPRM.

     The current  estimate  for  RD&T  is  based on  the same four
tasks  listed above,   as well as  on one-half   of  the cost  of
applying variable  injection  timing  (VIT).    The other half of
the cost for  developing VIT  is  included  in the RD&T costs for
NOx,  as  is half of  the cost of  improved turbocharging.   Thus,
the  particulate  standard is  being  allotted the full RD&T .cost
for  three  tasks  —  piston  modification,  transient  air/fuel
ratio control, and improvement  in  injectors —  as well as half
the  cost  for each of  the  two tasks of  improved  turbocharging
and  VIT.   The  cost  per  task  in  the   present  analysis  is
increased                        from                        the


  original  estimate  by  25  percent  in order  to  account  for  an
  additional  effort   to  optimize,  fuel  economy,  in  response  to
  comments  that  such   an  effort   will   occur.    The   estimated
  non-certification  RD&T cost  for  the 1988  particulate  standard
  is therefore  $3,292,000 per  task X (3  +  2(1/2))  tasks X  1.25
  fuel economy effort factor or about $16.5 million.

       In the Draft  RIA, 1988 HDDE certification was estimated to
  cost $13  million.   Assigning this  cost  at 50 percent  each for
  NOx  and particulate  allots $6.5  million  of  the  certification
 .costs   to  particulate  control.    This   brings   the   total
  undiscounted RD&T cost  to approximately $23 million.

       ii.   Variable Cost

       Hardware  costs,   like  fixed  costs,  are calculated  much as
  in the Draft RIA, where they were  estimated  at  $20  per  affected
  engine,  based  on   $5  per   modified  component.    The  modified
•--components   include:    1)   combustion  chamber/piston   design
  changes,  2)  injector  and  injection  pressure  modifications,  3)
  fuel    delivery   system    changes,    and    4)    turbocharger
  improvements.   This  analysis  has changed  only  to reflect the
  changes discussed above.

       The  cost  remains at $5  per  component for  the first  three
  components   listed   above,   while  the   charge   for   improved
  turbocharging  is  halved to  $2.50,  with the  other $2.50  being
  allotted  to  the cost of the  1988  NOx standard.   In response to
,  limited comments in this area,  an additional  $5  is  included for
  improvements in transient  control of air/fuel ratio control and
•  turbocharger operation.   As was mentioned  in  the  cost  analysis
-" for  the  1988  NOx   standard,  electronically  controlled  variable
  injection  timing  will  be used  to control  NOx  and  particulate.
  Adding  this capability is  expected  to cost  $25  per  engine.
  Half  of the cost  of  variable  injection timing,  or $12.50 per
  engine,  is  also  now charged  to  the  particulate  standard  to
  accompany  such  a   charge  being   added  to  the  NOx  standard.
  Summing  these  costs  results  in  a  $35 cost  for  each  engine
  receiving  the  modifications;  overall,  half  are  expected  to
  receive them.   The  average  cost per HDDE is therefore about $18.

       iii.  Total Manufacturer Cost

       It   is   expected  that   the  RD&T  costs   for   the   1988
  particulate  standard  will  be  incurred  according  to   the  time
  table   shown   in  Table 3-19,  which  is  proportional  to  that
  presented  in  the Draft RIA.  Costs are  shown  both undiscounted
  and  discounted to  1988 at  10 percent.   Total hardware costs for
  the  three model  years following  introduction of  the  standard
  are  based on projected sales for  those  years  as shown  in  Table
  3-11;  these  costs  are estimated  using  projected  sales figures
  and  are given  in Table 3-20  in both  undiscounted and discounted


                                    Table 3-19

                     -  HDDS RD&T Costs  for 1988 Particulate

      	Undiscounted	Discounted*	
          Non-                              Non-~
      Certification Certification   Total  Certification Certification   Total

1986    $15,OOOK       $1,OOOK     $16,OOOK   $18,150K       $1,210K    $19,360K

1987     $1,500K       $5,500K      $7,OOOK    $1,650K  '     $6,050K     $7,700K

TOTAL   $1S,500K       $6,500K     $23,OOOK   4l9,800K       $7,260K    $27,060K
     Discounted at 10 percent to 1988.


                     Table  3-20

      HDDE  Hardware CJSts "Eoc  1933  Pacticula-a











Discounted to  10 percent to  1988.


     Total manufacturer cost is the sum of  these  total  RD&T and
hardware  costs,  and  amounts   co  approximately  $41.5  million
undiscounted  and  $43.9  million discounted  cost,  as   shown  in
Table 3-21.

     b.    Cost to Users

     i.    First Cost

     The  total  RD&T  cost  developed  above  can be  recovered by
increasing HDDE  prices by  $28  for model year  1988-90  engines.
When added to  the  average  hardware cost of $18, the total first
price  increase  averages $46 per  HDDE.   Apportioning this  cost
only over  those vehicles affected by the  standard  results  in a
first price increase of about $84 per HDDE.

     ii .   Operating Cost

     As  described  in  the  technological  feasibility   analysis,
neither fuel  economy nor maintenance is expected to be impacted
by  the  0.6 g/BHP-hr  particulate  standard,  and hence  will not
impact user costs.

     iii.  Total User Cost

     The   average   increase  in  user  cost  due  to   the  1988
particulate standard is the sum of the  first price increase and
any  increase  in  operating  costs.   Operating  costs  are  not
expected  to  change,  so the  average user  cost  is simply the
first  cost  increase  of $46   per  model  year 1988-90 vehicle
equipped with an HDDE.           «•

     3.    Total Manufacturer  and User Costs  for 1988  NOx and
           Particulate Standards

     The  total  HDDE  manufacturer  cost  of  compliance  with the
1988  standards  is developed  above for  the  NOx and particulate
standards  separately.   These costs  are  shown together  in Table
3-22,  and total  approximately $110  million  manufacturer  cost
discounted to 1988.

     The  total  HDDE user  cost per  vehicle is  also   developed
above  separately  for the two  standards;  the total  is  shown in
Table  3-23  and is  $115-$810, depending  on  the  fuel economy
penalty.   This  value  will  tend  towards  $115  in  later  model
years as  fuel economy  improves.


                      Taole 3-21

     HDDE Manufacturer Costs for 1983 "articulate
	 	 1989 -•


RD&T Hardware
$27,060K $16,838K
Discounted to 10 percent to 1988.


                           Table 3-22

                 Total HDDE Manufacturer Costs
               1988 MOx and Particulate Standards



Grand Total        $105.871K
*    Discounted at 10 percent to 1988
** '  Model year 1988-90 HDDVs.

                              3-50 .

                           Table 3-23

                     Total HDDE User Costs
               1988 NOx and Particulate Standards

                                    Fleetwide Vehicle Average
                                  First Cost        Fuel Economy

NOx          "                         $ 69          $0-696
Particulate                            $ 46          $0
Total                                  4115          $0-696
Grand Total                          •      $115-810
     -The  $696  fuel economy cost  is  from a short-term 2 percent
     fuel economy  penalty.

           1991 NOx
Cost to HDDE Manufacturers
     The manufacturer  comments  which  applied  to  the  6.0  HDDE
NOx standard  generally applied  to  the originally  proposed  4.0
standard also, with  the  basic  assertion that EPA cost estimates
were too  low.   Specific values  for the  manufacturer  costs  of
meeting  the  lower  standard were   given  only  by  Ford  and  the
Department of Energy (DOE), as  follows:
Retail Price
Increase Per HDDE, 4.
NOx Standard
with new
          $350     from 6.0 to "lowest possible NOx"
          $700     from 10.7 to "lowest possible NOx"

          $643     from 10.7 to 4.0
     As  in  the analysis  for  the 1988  standard,  it  is unclear
what  is  included  in  the  cost  estimates  presented  in  the
comments in  regard  to  such things as RD&T,  markup,  and percent
of  engines  over  which costs  are apportioned.   The technology
changes which  are being used  to estimate these  costs  are also
not  detailed  in  the  comments,   although  Ford states  that all
engine   models   will   require   air-to-air  aftercooling.    And
finally, DOE presents  its  value,  developed  by Argonne National
Laboratories (ANL), as the total  cost of  going from the current
NOx  level  to  the  final  proposed  level,   rather  than as the
incremental  costs  involved with  the intermediate  level,  as EPA
does.   Ford  presents  costs  for  both the  total  and incremental
reductions,  but  finds  the costs  to  achieve  the  total  reduction
without  any  discounting  of  fixed  costs;  DOE  also  does not
discount,  making  it difficult  to  directly  compare  with   EPA's
     However,  the
those  projected  by
technology  and  using
standard.   The  larger
include  the total cost
approximately  twice  as  high
presumably  due to the  cost  of
         costs  given  in  the comments  are  close  to
         EPA,  when apportioned over  engines with  new
             incremental cost  from  the  intermediate
              values   given  by  Ford  and  DOE  which
             of controlling  from  10.7 to 4.0 NOx  are
                     as   EPA's  incremental   estimate,
                     the intermediate standard, which
was addressed above.

     Therefore, the analysis  of  HDDE manufacturer costs for  the
5.0  NOx standard  remains essentially  the same  as  that in  the
Draft  RIA  for  the 4.0  standard,  with  changes  only in  hardware
costs  in  order  to reflect  comments and  to  complement changes
made for the intermediate standard.


     i.     Fixed Cost

     RD&T costs for the proposed  4.0  g/BHP-hr  NOx standard were
developed in  the  Draft RIA.  The RD&T costs for  a 5.0 g/BHP-hr
NOx  standard  are  essentially  the  same  as   for  the  proposed
standard  at  $28,700,000 undiscounted cost, but  are  delayed one
year, along with the standard.

     ii.   Variable Cost

     Hardware  costs  applicable   to  the   5.0   NOx  standard  for
HDDEs accrue* from additional  and improved aftercooling,  piston
design  and  turbocharging.    The  Draft RIA also  included  costs
to  cover some portion  of  the   costs for applying  electronic
control modules  (ECMs).  However,-since  manufacturers' comments
have  indicated that virtually all  engines  will  have such units
for   reasons   other   than   emission    reductions   prior   to
implementation  of  the standard,  costs   for electronic  control
modules are  not  properly  attributable to  this standard.   Thus,
the  total hardware  cost estimates are reduced from those in the
Draft RIA.  The other  component costs remain  the same, however,
and  are  based on  comparisons to  costs  of   similar  pieces  of
equipment on existing  engines.

     Based   on   manufacturers'    comments,    engines   applying
aftercooling   for   the  first  time  are  most  likely  to .use
air-to-air  aftercooling.   As  discussed   in the  Draft RIA,  the
application  of  air-to-air  aftercooling is  estimated to  cost
$134.   It was shown  in the Draft  RIA   that 21  percent  of  all
HDDEs are turbocharged and employ no aftercooling;  10  percent
were  allocated funds  for  applying  aftercooling  in  response  to
the   1988   standard,   leaving   11  percent  still   without
aftercooling.  Applying the $134  per engine to  this  11  percent
results   in  an  average  of  $15  per HDDE.   The  turbocharged
engines which  did  not receive new or improved aftercooling for
the  intermediate  standard  will require   it now, at a cost of $73
for  converting  from  air-to-liquid  and   $91   for  upgrading  an
existing  air-to-liquid system.   These costs are taken from the
Draft RIA and  are the same  as used  above for  the 6.0 standard.
For  the 6.0  NOx standard it was  projected that 5 percent of all
HDDEs   would   convert   from    air-to-liquid  to    air-to-air
aftercooling and  16 percent would upgrade current air-to-liquid
systems.  For  the  1991 standard  it  is projected  that  another  5
percent  will convert  to air-to-air  and  15 percent will upgrade
air-to-liquid  systems.  On  a  weighted  basis,  the  average cost
of improved  aftercooling is  $17 per HDDE.

     The  component  cost  and  amount of application  of  piston
redesign  remains  as  in the Draft RIA, at  $5 per  engine  at a 25
percent   application  rate.   EPA's  best  estimate  based  on
manufacturer comments  results  in  $1.25 per HDDE.


     Earlier in this  analysis,  it was projected  that  50 percent
of the  HDDEs  would need  turbocharger  improvements to  ree-  tne
1988 NOx and  particulate  standards.   Improved  turbocharging  is
now  expected   to  be  employed  on  the   remaining  half of  the
engine's to  meet the  1991 NOx  standard  at  a  cost  of $5  per
engine,  as   allotted  for   the   intermediate   standard.   This
results in $2.50 per  HDDE.

     EGR is eliminated as an expenditure  to  meet  the  1991  NOx
standard,   in  a response  to  indications from  the manufacturers
that they  will not  employ  EGR  on  their  engines.   Electronic
control module hardware costs  are also  eliminated, as discussed
above,  although RD&T  costs were allocated for software design.

     Total hardware costs  are then the  total of  the above costs
for  aftercooling,  piston  design,  and   improved  turbocharging.
This amounts  to $36  per   engine  average hardware  cost,  or $113
per HDDE receiving  the new hardware.

     iii.   Total Manufacturer Cost

     To calculate  total  manufacturer  cost,  RD&T  and  hardware
costs  must  be  discounted to  the year   of  the  standard,  1991.
The distribution of RD&T  costs which was given in the Draft RIA
and is  used again  here  is shown  in Table  3-24.   Hardware costs
expended  according  to sales  projections and  discounted  to  the
year  of the  standard  are  shown  in   Table   3-25.   The  total
manufacturer  cost  arises   from the sum of these  costs,  and is
developed  in  undiscounted and discounted  forms in Table 3-26.
Total  manufacturer cost   of  compliance   with  the  1991  HDDE  NOx
standard is shown  to  be about $71 million undiscounted and  $73
million discounted  tc 1991

     b.    Cost to  Users

     i.    First Cost

     Incremental increases  in  first  cost due to  the  5.0 HDDE
NOx  standard  are  determined in  the  same manner  as  described
previously, except fixed  costs  are  recovered  over 1991 through
1993 model  year  HDDEs.   The average 'increase  in  first cost of
HDDEs  would consist   of  the  sum of the discounted  RD&T costs
amortized over  sales  for  model year 1991 through  1993  plus  the
hardware  cost developed  above.   These  costs  are approximately
$32  for RD&T  and  $36  for hardware  for a total  of  $68 average
HDDE  first  price  increase.  These "costs can  also be expressed
per  HDDE  requiring new technology  rather than as  average  per
engine  cost  by adding RD&T cost  apportioned only  over these
engines to  the  cost  of  the hardware required.   These  costs  are
approximately  $44  for RD&T  and  $113 for  hardware,  for a total
for $157 for  an engine receiving  all of  the new technology.


                             Table 3-24

                    HDDE aD&T Costs for  199L NOx

' 1.
Costs Cert. Costs Total
200K - 5.500K
200K $6,500K
. Costs

Cert. Costs Total
1,210K 18,
6,050K ' 7,
$7,260K $34,
Discounted at 10 percent  to  1991.


                           Tacle  3-25

                HDDE  Hardware  Costs  for  1991  NOx
14f 423K
     Sales taken from Table 3-11.
**   Discounted at 10 percent to 1991.


                       Table  3-26
	 	 . -_ -lg91
--. - 1993
RD&T Hardware
- 6,700K
- $13,673K
S28.700K $42,305K
RD&T Hardware
$34,837K $38,510K
 Discounted at  10 percent to 1991.


     The cumulative  increase  in  first  cost over current  engines
 to  achieve compliance with  the  1991 standard  would be  the  SU.TI
 of  the costs  for  1988  and  1991.   The  total increase  in  first
 cost i.s thus $69 for  1988 hardware  plus $68  for  1991 or  $137.

     ii.   Fuel Economy

     In  Chapter  2,  Technological  Feasibility,  it was estimated
 that  fuel  economy penalties  associated with the  1991  standard
 would  be  in  the  range of 0  to  1 percent  in the short term,  and
 this  penalty should  tend  to decrease  to  one-half  percent with
 time  as vehicle and  engine  improvements  are  made.   The 0 to  2
 percent  penalty  associated with the 1988 standard  should have
•disappeared  by 1991.          	         -   - -

     Fleetwide  fuel   economy  costs  are  calculated  in  the same
 manner as  for the  1988  standard,  amounting to an  average  per
 vehicle  lifetime  increase   of  $348 per  1  percent  increase  in
 fuel consumption.  With a 0  to  1 percent  change in  fuel  economy
 expected   for  the   5.0  standard,   the  short   term  fuel  cost
 increase  is  thus  $0  to  $348, tapering  off  to  $0 to $174 over
 the long term.

     Several  commenters—Mack,  American  Trucking  Association,
 and Department  of  Energy—indicated  the  fuel cost increases
 would  be much greater than  this.   All estimates, however, were
 based  on a  variety  of  different  assumptions  regarding  vehicle
 lifetimes  and  amount of  fuel  currently used,   as  well as  on
 higher fuel  economy  penalties.   The fuel  economy penalty  issue
 is  most important, and is addressed in Chapter 2, Technological
 Feasibility;  the  other  issues  are methodology  differences  of
 VMT and fuel price estimates, and  are  discussed above in regard
 to  the 1988  HDDE NOx standard.

      i i i.  Maintenance

     Maintenance   is  not  expected  to  be  affected   by   this
•standard,  and  hence  should not  impact on cost.

      iv.   Total User Cost

      In summary,  owners of  model  year  1991 and  later  vehicles
 which  are  equipped with HDDEs can  be expected to  pay  an average
 of  approximately  $68 incrementally over model  year  1988 through
 1990  vehicle  prices  or  $137  total more  than  they would  have
 paid  without  the  introduction  of  "the two  HDDE  NOx  standards.
 In  terms of  fuel costs,  the  increased average  lifetime  cost  per
 vehicle is expected  to be between  $0  and $348, tapering off in
 later  model  years  to $0  to $174.   Incremental  lifetime  increase
 is  thus $68  to  $416 in the short  term,  and $68  to  $242 in  the
 long   term.   Total lifetime  increase  for model  year  1991  and


later  HDDEs  due  to the  NOx standards  is  $137 to $485  in the
short term, and $137 to $311 in the  long  term.   These costs are
summarized in Table 3-27.

     5.    1991 Diesel  Particulate  Standards  (0.25 g/BHP-hr for
           HDDEs with 0.10 g/BHP-hr for Urban Buses)

     In this  section,  the costs of the  1991  diesel particulate
standards  for   HDOEs  are  examined.    As   described  in  the
Technical  Feasibility  Chapter  (Chapter  2),  achieving  these
standards  will  require  the  use of  trap-oxidizer  technology on
100 percent of  the urban buses and  about 60-70 percent  of .the
remaining  HDDEs.    Since  the same  basic type  of  trap-oxidizer
system  will  be  used  on  both  HDDEs  and urban  buses,   in  the
subsequent  analysis  of  comments  and  cost  derivations,  the
primary discussion  in each section  centers  on HDDEs in general,
and   is   then   followed  by   a   discussion   of   any  special
considerations of urban buses,  as necessary.

 •	a.   - Cost to  Manufacturers   — ~.  _   .   ._. .

     i.    Fixed Cost

     In   the   draft  analysis,-  EPA  separated   research  and
development  costs  into  three  categories:   1) general  system
development;  2)  specific  engine  family  design;- 3)  electronic
control development.   The seven largest HDDE manufacturers were
each  allotted  about  $2.8  million   to  develop  general  trap
systems.   Smaller   manufacturers   were  expected  to  rely  on
guidance  from trap-oxidizer manufacturers  for  general designs.
Engine  family specific designs were assumed to  be required by
about  70  percent "of  the  engine families  with  averaging,  at  a
cost  of about $230,000  per  engine  family.   The  development of
electronic  controls was  estimated  at  about  $115,000  for each
engine  family.

     Three comments were received  regarding  EPA's research and
development  cost estimates.   In  the  first  comment,  GM stated
that  EPA  had clearly  underestimated  the  cost  of  basic  trap
development,  claiming  that  it  had already  expended $40 million
by   the   end   of   1984.    In  presenting   it's   $40  million
expenditure,  GM  failed  to  distinguish  what  portion of  this
amount  is  attributable  to  LDD trap  development   and  which is
attributable   to    HDDE   trap  development.     Without   this
information,  it  is impossible to  know how  much  GM  has indeed
spent  on  trap systems  for  HDDEs.   The  company's  claim  can be
placed   in   perspective,   however,    by   the  fact   that  LDD
regulations  requiring  trap  technology were  promulgated  for the
1985  model  year  in  early  1980.[4]    (These requirements were
subsequently  delayed   in  early   1984  until  the  1987  model
year.) [5]   By  contrast,  HDDE  trap  requirements   for  the 1991
model  year are  just now being  promulgated  in this rulemaking.


                           Table 3-27

                   Heavy Duty Diesel Engines
     Discounted* User Cost  set Engine for I9S8 and 1991 NOx

                 1988 Standard
 $0 to 696
$69 to 765
                1991 Standard
 $0 to 348
$68 to 416
  $0 to 348
$137 to 485
    $ 0
$0 to $174
$68 to 242
$0 to 174  •
$137 to 311
     At  10  percent  per year  to the  year  that standard becomes


 Additionally,  most trap-related technical  information  submitted
 to   EPA   by   GM  has   concerned  light-duty  traps.    It   seerr.s
 reasonable  to  expect,  therefore,  that only  a  small portion  of
 GM's total trap  development  expenditures  should  be  attributed
 to   the  HDDE  trap standards.   The  Agency would  also like  to
 point  out  that  due   to  the  problematic   nature  of  estimating
 development    expenditures   for   each    manufacturer,    EPA's
 projection  should be  reviewed in  terms  of  an  overall  average
 per  manufacturer,  with  some  spending more and  others  less.
 Because  of  its size,   it  is  not unreasonable  to  expect that  GM
 should   fall   into  the  former  category.   Hence,  GM's  comment
 provides no   basis  for  revising  EPA's   original   estimate  of
 general  development costs.

     The second   comment,  which  was  confidential,  identified
 another  company's expenditures  for HDDE  trap  development  from
 1979 through  1984.  In this  case,  the reported values  were well
 within   EPA's  estimate for  each manufacturer.   Therefore,  the
 original estimate for general system development  appears  to  be
 appropriate  based on  this comment.

     The third comment came  from International Harvester  which
 claimed  that  its mechanical  durability testing would require 11
 LHDDVs   and 13 MHDDVs  with each  vehicle  successfully  traveling
 10&,000   miles  and   150,000   miles,   respectively.   While  not
 agreeing with the  need  for  such  a  large  test  fleet,  EPA
 calculates  the  cost   of  such a program  at about  $2.3  million.
 This amount  is less  than  half of the development cost  for  IH' as
 derived  in the draft  analysis.  Therefore, the comment provides
 no  basis for  changing EPA's  original  research  and  development
.estimates.           ,

     Another   area  of  comment  concerning   development  costs was
 that of vehicle  modifications.    The  draft  analysis  did  not
 contain   a  cost for development  and tooling expenditures  which
 might  be needed to modify the vehicle assembly to  integrate the
 trap-oxidizer into the overall design.   International  Harvester
 expressed  the   strongest   concerns  regarding  the  potential
 magnitude of  vehicle modifications.   In   an  apparent  reference
 to.  tractor-trailer  combinations,   IH  stated that  if  two  traps
 are necessary and  their size  requires  that they be  mounted
 behind the cab, then  the  location of such  things as  the sleeper
 unit,  fuel tank,  air tanks,  fuel  and oil  filters,  aerodynamic
 side  shields,  etc.   may  be  affected.    This   would  in  turn
 adversely  affect hundreds   of  body  builders.   General  Motors
 stated  that  in  its  evaluation  of  possible   vehicle  design
 changes, there  appears  to be  adequate  room within  the vehicle
 frame  on  a  MHDDV to  mount  a trap  and  muffler,  although  a few
 vehicle  components may  need to be relocated on some  versions.
 For its  HHDVs, GM claimed that if mounted  behind the cab,  traps
 may restrict  the vehicles  turning  radius which,   in  turn,  may
 reduce  tractor-trailer offerings.  General Motors  also alledged


 that  "essentially"  no space  was available  for  a trap  in it's
 urban bus, and  that  significant  redesign of this  vehicle would
 be  required.   The company  claimed  that  potential  changes  may
 include  relocating  air  conditioner   and   heater  components,
 eliminating  seating  for  up  to  five passengers,  or installing
 new   suspension   systems   and   bulkheads.   Cummins  generally
 commented  on   the   need  for   vehicle  modifications  without
 identifying specific changes required.

      EPA agrees with  the commenters  to the  extent  that  vehicle
 modifications may be  required  for  certain  trucks  in  order to
 accommodate trap oxidizers.   The  extent  of  any required vehicle
 modifications will  obviously be dictated  by  -the type  of trap
 system  ultimately  chosen   by   manufacturers,   and  upon  its
 specific configuration.  Since  neither  the  final trap type nor
 specific configurations have yet  been  identified, the potential
 costs  associated with  required vehicle  redesign  can  not  be
-.quantified to  any degree.   However,  it is  possible to discuss
 in general terms  the  types  of trucks that are most  likely to be
 affected and how  the  negative effects  of any  redesign  might be

      As  stated  by  GM,   the  greatest  potential   for  vehicle
 modifications  is  associated  with  HHDDVs   and  urban  buses.
 Regarding  HHDDVs,  and all other  diesel  trucks for that matter,
 it  should  be  remembered  that  diesel  particulate  emissions
 averaging  will   result  in  a  significant  number of  trucks  not
 needing  traps.   To  the  extent   that  a   manufacturer  can
 anticipate  problem   installations,   such  vehicles/engines might
 be  excluded  from having traps.   Beyond this,  it is reasonable
 to  expect  that  many  HHDDVs  and  urban  buses will  normally
• undergo some design  changes by  the  1991 effective  date  of the
 standards, especially  in light  of the  emphasis  being placed on
 improved   aerodynamics  by   HDDV   manufacturers.    For  such
 vehicles,  the  incremental  cost   of  incorporating traps  in the
 design would be  minimal.   Finally,   there are  many vehicles for
 which  creative  packaging of  the trap  system  will  avoid costly
 redesigns.   For   example,   Southwest   Research  Institute  is
 currently  testing a GM  coach  engine with  a trap configured to
 replace the  engine's  exhaust  manifold.[6]   Such  a  design would
 require no redesign of the urban  bus.

      Overall,  then,  while  modifications will likely be needed
 on  some vehicles,  the extend  of these  changes will  in large
 part  be  dictated by  the  engine  manufacturers   choice  of trap
 system,  and  the  foresight  with which  it   is  configured  or
 packaged  to  meet the  requirements  of  the  vehicle  manufacturer
 or  body builder.  Therefore,  EPA believes  that the  number of
 significant  design  changes  can  be minimized,  and when averaged
 over  the fleet their impact will  be  small.   Because of this, no
 fixed  cost for  vehicle  modifications  will be  included  in the
 cost  of the regulations.


      In  addition  to the  expense  of  research  and development,
 the   draft  analysis   included,  the  fixed  cost  of  emission
 certification  testing.   No comments were  received  on this cost
 component,  so  it is being  retained  here  without change.

      In  summary,  none  of  the comments supported any changes  to
 the   fixed  costs  of the  draft  analysis.   They  are therefore
 being retained  unchanged.   Nonetheless,  it  is  interesting  to
 note   that  even   if  the  comments   had   provided   a  basis   for
 revising  the  fixed cost estimates, any corresponding change  in
 the total  cost of the regulations would  be very small.   As will
 become evident  later  in  this  analysis,  fixed costs  are only
 about eight percent of  the total  cost.

   -   As  shown  in  Table  3-28,  the  total  fixed cost of  the 1991
 particulate standards  is  $49.5  million.   The  distribution  of
 the expenditures over time is  the  same  as originally estimated
 in the draft  analysis,  except each allocation has been delayed
 one   year  to  account  for  the  revised  effective  date  of   the
 standards  from 1990 to  1991.

      ii.    Variable Costs

      The   draft  analysis   explored the  potential  use  of   two
 significantly  different  types of  trap  oxidizers*  for  HDDEs:  a
 non-catalyzed/  ceramic   monolith   system;   and   a  catalyzed
 wire-mesh  system.   As fully described in  the Diesel  Particul-ate
 Study  (DPS),[2]   which accompanied  the  draft  analysis,  both
 systems  function  similarly  in   that  particulate  matter   is
.filtered  from  the exhaust and then  periodically  burned in  the
 trap   to  prevent  excessive  exhaust  backpressures  which  would
•degrade  engine  performance  and fuel economy.  This  latter step
 is termed  "regeneration"  and  is  significantly  different in  the
 two   trap   types,   depending  on the  presence or   absence  of  a
 catalyst.   The  ceramic monolith  design  assumed   in  the  draft
 analysis  used a fuel burner  to  heat the trapped material  to  its
 ignition  point.   During this process,  the  engine  exhaust  flow
 is temporarily routed  around the  trap,  while the  burner  and
 trap   are   supplied with  a  controlled   air  supply  to   ensure
 adequate   oxidation of  the  trapped  material without excessive
 heating.   The  regeneration  system with  a  catalyzed wire-mesh
 trap   can   be   less  complex,  since  the  requisite  temperature
 increase  of the trap  is significantly less.   The  catalyst  trap
 evaluated   in  the  draft  analysis  was  assumed  to achieve ' the
 required   moderate  heat   rise   by  delayed  in-cylinder   fuel
 injection; thereby,  eliminating  the need for a fuel burner  and
 bypass system.

      In  assessing  the  variable or  hardware  cost of  the  two
 systems,   the  draft  analysis  found  that  the  wire-mesh  design
 with   its  catalyst coating  was  quite  expensive.   Hence,   the


          TaoLe 3-28

   Total Fixed  Costs o£ the
199 L HDDS Particulate Standards
vear Develooment
L w O fc
- 1990
_ _2
. OM
Certification Tot
Sl.OM 14
5.5M- _7
S6.5M $49
. OM


ceramic  monolith trap  with  a  fuel  burner  regeneration  system
was used  to  estimate the cost of the proposed  HDDE particulate
standards.   A  summary of  the  hardware  costs that  were  used in
the proposal are shown in Table 3-29.

     Comments  on   the   cost  of  trap-oxidizer   systems   were
received   from   six   manufacturers,    one   of   which   was
confidential.    Cost  figures   were  also   provided   by   the
Department  of  Transportation  and   the   Department  of  Energy.
Unfortunately, each of  the  government  estimates  were  supplied
as  a   composite'   total   of  the  trap-oxidizer  standards  in
conjunction  with either  the  1988 non-trap standard or  the  1991
HDDE  NOx  standard   and,  therefore,  could  not  be analyzed  in
detail.   Hence,  these  latter  two  comments  are  not  discussed
further.   The  non-confidential  industry  estimates are  displayed
in Table  3-30.

     The  manufacturers  estimates in Table  3-30  were  generally
reported   as   the   cost  of  a  total system.    Little,  if  any
information  was  provided as  to  the  derivation  or  basis  of the
estimates.   For  example,  the  manufacturers  usually provided no
breakdown   of  the  total   system   into   its   various   cost
components.   Also,   the  estimates were  variously  described as
"cost  to  the consumer"  or "consumer effect."   Therefore,  it is
unclear  if  some  of these  costs   include  fixed  or  operating
costs,  or whether  they  inappropriately  reflect  the full retail
cost  of a trap  system,  rather than the  incremental cost fo-r a
new vehicle.   Even  if only a  1  percent penalty  were included in
some  of these estimates,  this could add  about  $350 to the total
cost for  an  average HDDE when  discounted  to  the year  of vehicle

     Cummins  was  the only manufacturer  shown in Table 3-30 that
provided  a breakdown of its  system  cost  by  hardware  component.
As reported  by the  company,  the component costs  are as follows:

      1.    Trap  Substrate  Material            -  $720-$!,080
     2.    Trap  Casing and Ceramic Mounting       $250
     3.    Diesel Burner for Regeneration         $400
     4.    Electric Air  Blower  for Burner         $175
      5.    Miscellaneous Control  Costs            $650	

           Total                             $2,195-$2,555

     Cummins  describes  the   trap  substrate,  i.e.,   ceramic
monolith,  as  being  60-90   liters  for   a  "possible  dual  trap
option"   at    $12   per   liter.    Additionally,   the  costs  are
described  as  component  costs  from   suppliers,   without  an
allowance for assembly, machining of other  ancillary  parts, or
fixed  manufacturing costs.


Tacle 3-29
Trao-Oxidizer Va
Cost Category
Ceramic Monolith and Hous
Fuel Delivery System
Fuel Ignition System
Auxiliary Air System
Exhaust Diversion System
Total Hardware
ri=ble Cost
ing $207
Fror, the Draf1:
31 A


                           Table 3-~30

          Manufacturers'  Trap-Oxidizer Cost

 Source             Cose         	Description	

 IH              $1285-2070       MHDDE, single trap
                    4710         HHDDE, single trap
                    7210         HHDDE, dual trap if required

 Ford                2200         No comment

 Cunmins          2810-3270       HHDDE, possible dual trap option

 GA                575-900        LHDDE
                    2300         MHDDE, single trap
-	 -    -         4000         HHDDE, dual trap

 Saab               2500+         No comment


     The  confidential  comment  also  provided  some  breakdown  of
 cost  oy component.  Due  to  the confidentiality of the  comment,
 however,  all that can  be said  is  that in  contrast  to  Cummins
 estimate,  the  reported  trap cost, (i.e.,  substrate and  housing)
 is  substantially  less  expensive,   while  the  air   supply   and
 burner  are  somewhat more  expensive.
 Cummins',   makes   a   rigorous   analysis   of   the   estimates
 impossible.   The only  clear  conclusion that  can be reached  is
 that  the  cost  of  a trap-oxidizer  system as  reflected by  the
 manufacturers   comments  is  substantially  higher   than   EPA's
 estimates  from the  Draft  Analysis  (Table 3-29).   In  order  to
 address   this   disparity,   EPA's  only  recourse  has   been  to
 completely reevaluate  its  estimates,  using  the  comments  where
 possible,   to   better   define   the  variable   costs   of   the
 regulations.   In  preparing  this  reevaluation, EPA has  also made
-use  of  an  independent  contractor  to  prepare  component  cost

     Overview  of  the Analysis

     In  reevaluating the  variable  cost  of  trap-oxidizers  for
 HDDEs,  EPA will  examine  the component  costs of three  separate
 systems.   The  first system uses  a  ceramic monolith  trap  in
 conjunction   with   a  fuel  burner   and   exhaust   bypass  for
 regeneration.   This  is very similar  to the trap-oxidizer unit
 used   in   the   draft analysis   to   estimate  the  costs  of  the
 proposal.   The second system is  the same  as the first, in that
 a ceramic monolith  is  used to  filter the  exhaust,  but differs
 with   regard  to  the  type  of   heat  source  used   to   initiate
 regeneration.   In  this  system,  electric  heating  elements  are
 included   in  "the  trap  housing  and  are  energized  with  the
 vehicle's  electrical system. An  exhaust bypass  is  also required
 with  this  system.   Electrically  regenerating the  trap can  be
 advantageous    since   it   eliminates   the   bulk    and   safety
 considerations  associated with -the fuel  burner  approach.   The
 third  system  is  radically different  from the  others in both
 trap  design and  method of  regeneration.   The filter medium  in
 this  trap design  is composed  of ceramic  fibers  which  are  wound
 into   a  type  of  fabric.   Regenerating  the  ceramic   trap  is
 accomplished  through the use  of a  metallic catalyst  compound
 that  is  injected  into  the  exhaust  at  the time  regeneration  is.
 desired.   Since the catalyst substantially reduces  the ignition
 temperature of  the trap particulate material, no exhaust  bypass
 system  is  required.   This  system   is   attractive  primarily
 because  its simplicity could  result  in  reduced> costs  compared
 to the  other  two  systems.   Each  of these systems  will  be
 further  described below.


     Most  research  and  testing  to  date  have  focused  on  the
 ceramic   monolith   trap.    Of  the  two   regeneration   methods
 described  above,  the fuel burner concept  has been  successfully
 used  in  vehicle  tests  and  its hardware  components  are well
 understood.   The electrical  regeneration  system,   on  the  other
 hand,  is  much  less  well  defined  at  the present  time.   The
 ceramic  fiber  trap  and  catalytic   regeneration  system  is  the
 most   recent   entrant   in the  trap-oxidizer  field.   The trap
 concept   is   proprietary  to  Mercedes-Benz  AG  and  relatively
 little   is  known  about  it  compared  to   the  ceramic monolith
 trap.   Due to  the  present  state  of  knowledge,  the  cose of  a
 ceramic  monolith/fuel  burner system can be  estimated with  .the
 greatest degree  of certainty.   Therefore,   as   in  the  draft
 analysis,  this system will  be  used  to  derive the  variable costs
 associated   with    the   1991  particulate   standards.    Also,
 considering  its state of development,  this trap-oxidizer design
.could  be the  first commercially available system.

     The variable costs of the ceramic -monolith/electric system
 and   ceramic   fiber/catalyst  system   are  still   of   interest,
 however,  since their potential advantages  may result  in either
 or  both  of  these supplanting  the  ceramic monolith/fuel  burner
 design.   Therefore,  these systems are  examined  here  to  provide
 a measure of  sensitivity to  the overall cost  estimates.

     The variable cost of each  trap-oxidizer  system is found by
 determining the retail price equivalent (RPE) of  each component
 part.    With   few  exceptions,  the   costs  were  based on work
 performed by Mueller  Associates  under  contract  to EPA.[7,  8]
 The  contractor's estimates  were  based on  the  manufacturer's
 cost  of  each component.   To  obtain the  required RPE  of  the
"various  components, EPA  adjusted  the  contractor's  estimates to
 reflect   the   added  costs   associated  with  a   manufacturer's
 overhead and profit,  in addition to dealer  costs.  The  mark-up
 factor  used to derive the RPE of each component was  1.29 (i.e.,
 a  29  percent  increase) .   This  factor  has  been  used  in  past
 rulemaking  actions and is derived  in Reference 9.

      The costs  not  taken  directly from  the  contractor  were
 estimated  by  EPA  and  will be  specifically  identified were
 applicable in the discussion.  The  Agency's  estimates are based
 on   previous   work  by  Lindgren,[10]  information  supplied   by
 Mueller    Associates,[7,8]    the   DPS,    and   on   engineering
 evaluations of similar automotive  components.

      Now  that  the  general  methodology  has  been  described,
 estimates of  the  various  component costs  can  be  presented.
 This  will be  done  first  for the  two trap designs  and then  for
 the  three regeneration systems.  After  the component  parts have
 been estimated,  tne resulting  total cost  of  each  system  will be


       Ceramic  Monolith Trap  Costs  —  The cost  of any  specific
  trap design  is  dependent  on the volumetric  requirements  of  the
  filtering medium.   One of the  most  important  considerations  in
  sizing, a trap is to  make  it large enough to avoid undue exhaust
  backpressures,   which  would  degrade  engine   performance  and
  adversely affect fuel  economy.   In general, the  requisite trap
  volume  for   a  given  engine  is  dependent   on  the  volumetric
  exhaust  flow during  normal  operation.   The   HDDE  trap  sizes
  assumed in the proposal were  derived  in  the DPS.  The  analysis
  was  based  on  successful  testing  of a 5.0  liter  trap on  a
  Mercedes-Benz  300D   by   Southwest   Research   Institute   for
  EPA.[11]    Since   volumetric  exhaust   flow   from  an  engine  is
  roughly  a  function  of  the  amount  of   fuel- consumed  (i.e.,
  inverse  of   fuel  economy),  HDDE  trap sizes  were estimated  by
  increasing  the  LDDV trap  size (i.e., 5.0  liter) by  the  ratio  of
  the vehicle's MPG  (26  MPG)  and the projected  average  MPG's  of
  each  HDDE  size  category.   The resulting  HDDE  trap  volumes,
  which  were  used to  estimate the cost of  the  proposal,  varied
  from 8.3 to 18.4  liters depending  on HDDE size.

       No comments were  received on  this  method of  sizing  HDDE
 , traps.  The  Agency's  estimated trap  volumes,  however,  contrast
  sharply with cost  comment  from Cummins indicating the  use  of  a
  60-90  liter  trap  in  its calculations.   Unfortunately, Cummins
  provided  no  information  describing   how   it  arrived  at  this
  size.   In  addition,  trap  test data  supplied  by GM (described
  further  below)  are  based  upon  trap  volumes   somewhat greater
,v than  estimated  by  EPA.    This disparity   has  caused EPA  to
,, reevaluate its  sizing methodology.

       The major  difficulty  in  attempting  to   estimate  exhaust
  flow  changes from  one  vehicle  or  engine type  tc another is  in
  choosing a sizing  parameter that accurately reflects  the  many
  variables  which   ultimately   determine  the   actual   exhaust
  volume.  Such key variables include air/fuel  ratio,  vehicle and
  engine speed, engine efficiencies,  and how  the engine  is loaded
  in  normal  operation,  i.e.,  what   percentage   of the  engine's
  rated  horsepower  is typically  used.   As  vehicles  become more
  disparate  in size  and  function,   the accuracy  of   any  single
  parameter  for  estimating exhaust volumes will diminish  due  to
  the multitude  of  variable  interactions.  This notwithstanding,
  EPA continues  to  believe that  fuel consumption is a reasonable
  surrogate  for  approximating   exhaust  flows.   It  inherently
  accounts for many  of the  changes  in  vehicle  operating  regimes
  and engine operating parameters among  vehicle types.

       While fuel  consumption appears  to  be  a  useful  method, for
  estimating exhaust  flows,  EPA  also   recognizes  that  the wide
  disparity in operating regimes  of  LDDV engines  and  some HDDEs,
  especially the  heaviest trucks,  could  result  in this  approach
  being  somewhat inaccurate  for  such  HDDEs.   Therefore,  at least
  for  some  applications,  engine  horsepower  also may  be  a useful
  sizing parameter.


      To explore this method,  a  review  of EPA trap-oxidizer test
 programs  and  the  information  submitted  in  response   to  the.
 proposal  was  conducted.    Thi's   review  showed  that   several
 different  combinations  of  trap  volume  and engine  horsepower
 have  been tested.   Two EPA  test programs  are useful  in this
 comparison.   First,   the  previously  referenced  Mercedes-Benz
 300D  test was  conducted  with  a  5.0  liter  ceramic  monolith
 trap.  The rated  power of the engine  used in this vehicle model
 is  118  HP. [12]   Second,  EPA has  also  tested  a GM  bus  engine
 with  a 20 liter  trap. [13]  This MHDDE has  a rated power of 190
 HP.[12]   While  -the   results  were   quite   variable,    overall
 essentially  no fuel economy penalty was observed with this trap
. volume/engine  size combination.   The  trap  volume  (liters)  to
 horsepower   ratios  from  these  tests  are   about  0.04  for  the
 "EPA-LDDV" and about  0.10  for the "EPA-MHDDE."

      The  comments  contained  test   data  for  two  additional
 engines.  The  first engine has a  rating of  205 HP, [12]   and was
 tested by GM with  trap sizes ranging from  20-25  liters.  The
 information  submitted by  GM  for  some  of these tests shows what
 appears  to  be quite  reasonable  pressure drops  across  the trap
 during actual  over-the-road  vehicle  testing.  Hence, this trap
 volume/engine  combination may   represent   an   acceptable  trap
 volume  to horsepower  ratio  from  the standpoint  of minimizing
 any potential fuel  economy penalty.   The average trap volume to
 horsepower ratio  for  the  "GM-MHDDE" is  0.11.   The comments also
. contain  confidential   test  results  on  a LHDDE  vehicle.   Due to
 the confidential  nature  of  the  comment, however,  all  that 'can
 be  stated is that  the  trap volume  to  horsepower ratio for this
 test  was somewhat  less  than  the EPA-MHDDE factor.  The  Cummins
 comment  regarding  trap volumes is  not  used here since the basis
 of  the estimate was not  reported.

      Table 3-31  presents  a  summary  of the average trap  volumes
 that  result  from  applying the various  factors  discussed above
 (i.e.,  both  MPG   and  horsepower  based)  to the  average fuel
 economy  and  horsepower  ratings  for  the  various  HDDEs.  Note
 that  the MPG values  shown in the  table have been updated from
 those  in the DPS,  as  discussed  in  the section on the 1991 HDDE
 NOx standard.   This  has resulted in  revised HDDE trap  volumes
 using the fuel consumption sizing method.

       From the table,  it  is  readily   apparent  that  the  various
 approaches   result  in  a wide  range  of  estimates.   The  lowest
 values  are  consistently estimated  by  the  EPA-LDDV horsepower
 approach.    This  is   not  surprising  given   that  a  trap would
 likely  be  sized  or   optimized  for the most  typical  type  of
 operation.   The   power  requirements  of  a  LDDV  under   normal
 operation is usually less  than  that  for  a diesel  truck when
 expressed as  a   percentage  of  the engine's rated  horsepower.
 For  LHDDEs   the   difference   may  be   rather   small,  but   the



                       of HPEE Trap Volume Esrimacea
                             Fuel Economy and Horsepower Based Volumes
THTTW      15 1      130      8.6L
«       K     'S     SS
14. OL
13. OL
18. 5L
35. OL
14. 4L


disparity  becomes  progressively  greater  as  the  HDDE  size  is
increased.  As  a result,  the  EPA-LDDV horsepower  factor  would
progressively  underestimate HDDE  trap  volume requires as  the
HDDE size of the vehicle increased.

     At  this  point, some  judgment  must be  used  in conjunction
with  the  remaining  fuel   economy  based  and horsepower  based
values  shown  in  Table 3-31  to identify  reasonable HDDE  trap
volumes.  An extremely important  consideration in this decision
is  the  tradeoff   between  trap  volume  and  effects  on  fuel
economy.   In  this  respect,  it  can generally  be  concluded that
the cost of a  somewhat larger  trap  is  much less  than  the c.ost
of adversely affecting fuel economy by using  a trap that is too
small.  Hence, it is EPA's  intent  to  be conservative,  (i.e.,  to
err . toward   larger   volumes)  in   estimating   trap   volume

     Most  LHDDEs are  loaded  and  driven much  like LDDVs.   This
argues  strongly  in   favor  of  using  the  fuel   economy  based
estimate,  since  this  method  should be  quite accurate  in this
case.   To be  conservative,  however,  the  trap  size  for  this
category  will be  estimated  by  averaging  the   EPA LDDV  fuel
economy   based  volume   with  the   various  horsepower  based
volumes.   The  result  is  an estimated volume  of  11  liters  for
LHDDEs.   Considering  that  the  operating  regimes-of MHDDEs  and
HHDDEs  become  increasingly  dissimilar to  LDDVs  and  LHDDEs  as
truck  size grows,  and that  the  EPA and  GM horsepower factors
are based on tests  that should  not  result  in undue fuel economy
penalties,  the  GM-MHDDE  horsepower  factor  will  be   used  to
estimate  trap   volumes   for   these  vehicles.   The  resulting
estimates are 21  liters for MHDDE and 39 liters for HHDDEs.

     Another  important detail  which  must be  dealt with before
the trap costs can  be  estimated is  the  number  of traps that may
be  required  by the  variously sized  HDDEs.   The  "DPS assumed the
number of traps  for each HDDE size  as follows:  one for LHDDEs,
two  for  MHDDEs,  and two  for  HHDDEs.   No comments were  received
regarding the number of traps for  LHDDEs.   The comments from IH
and GM indicate  one  trap  will be  sufficient  for MHDDEs  (Table
3-30).  GM provided an illustration  of  this concept that showed
two  ceramic   monoliths  arranged  in  series  to   provide  the
necessary volume.

     The comments  for  HHDDEs are indecisive with  regard to- the
number  of  traps  per  vehicle (Table  3-30) .   The Cummins and IH
comments  suggest single traps  are possible.  Also  the Cummins
comment  regarding  the  possible  dual trap  requirement would seem
to  be  invalidated  given  that  the  trap   volume   requirements
estimated  above  are  significantly  less   than  assumed  in  its
description.   GM's  comment was provided in the  context of  its
12.1  liter, 435  HP engine.  This engine is  significantly  larger


than post  of  the engines  in this  HDDE size  category.   It  is
possible such a  large engine  -night  require two traps.  However,
even in  this situation,  placing trap  monoliths  in  series  to
attain, the  required  volume is possible, although  some  increase
in bacKpressure  would  result.   Based  on  the  comments   and  the
requisite trap  volumes,  EPA  believes  that for most  if not  all
HDDEs,  a single  trap  is  sufficient if  not  preferable  for system

     The ceramic  monolith  trap  has several components.   These
components   include  the  monolith itself,  a  ceramic  mat  and  a
trap  housing.    Each component   and  its   associated  cost  is
discussed below.

     The ceramic  monolith material  is  constructed as  a  matrix
of alternatively opened and closed  cells.   Particulate material
is collected as the  exhaust  flows  through  the  porous  wall  of
one  channel  into the  next.  The monoliths  used  for HDDEs  are
assumed  to  be   approximately 12  inches in  diameter,  although
smaller sizes can also  be made.   The cost  is estimated at about
$6  per  liter,   based  on  information from  Corning,  one  of  the
largest  ceramic   monolith   manufacturers.[8]    This   can   be
contrasted  with the  $12  per liter  cited  by Cummins  in  its
comment, which  was  unreferenced.   Using the  former  value,  the
estimated  price  of   the  ceramic  monolith  material  for  the
various  HDDEs  is:   $66  for  LHDDEs,  $126  for MHDDEs,  and $234
for HHDDEs.

     The ceramic  mat holds  the monolith   securely  within  the
housing.   It  also functions  as   a  shock  absorber and  provides
thermal insulation.   This  item is estimated at $3, $6,  and  $12
for LHDDEs. MHDDEs,  and  HHDDEs,   respectively.[8]

     The trap housing encloses the  ceramic monolith  and  ceramic
mat.   It  includes  baffles,  flanges,  and pipe  connectors (used
to  connect  the  trap  to the  exhaust  system and fuel burner or
bypass  valve),  in addition to  fittings for  mounting  sensors.
The  estimated cost  is  $32  for  LHDDEs,  $40  for  MHDDEs,  and $46
for HHDDEs.[8]

     Table 3-32  presents  the  total  estimated cost  of a  ceramic
monolith trap for each of the HDDE size categories.

     Ceramic  Fiber   Trap  Costs   —  Other   than  the  basic
construction of  this trap  design,  little   specific  information
is available.   In general,  perforated stainless steel  cylinders
are wound with  silica fibers until the  desired  filter  "fabric"
has  been  created.   Several  such  tubes  are  then arranged  in
parallel inside the stainless steel  housing  so that  the  exhaust
must  pass  through the  fabric and  into the  stainless  cylinder
before ' exiting  the  trap.   To estimate the  cost   of  this trap
design, the  volumetric  requirements  that were developed  for the


                           Table 3-32

                    P^Mmated HD-DE Trap Costs

Tr.n Design               .   LHDDE         _MHDDE_        _HHDDE.

Ceramic Monolith             $101           $1"           $292
Ceramic Fiber                 73            106


ceramic monolith are assurred co apply.  The cost  for  the entire
cera-ic trap is estinated  at  573  for LHDDE,  $106 for KKD3E, and
$140 for HHDDE.[7]   These costs are also shown in Table 3-32.

     Fuel Burner Regeneration  System  Costs  -- A  typical system
of  this  type  has  several  primary components:  a  burner can,  a
fuel  delivery  system,  an  ignition  system,  an  auxiliary  air
supply  system,  an exhaust  diversion  system,  and  an  electronic
control  system.    An  additional   component   required  by  the
trap-oxidizer which  is  actually  neither part  of the  trap nor
regeneration system is  the  exhaust  pipe.   The  cost  of  this
component and the  others  are discussed below.

     The burner can is located just  upstream of  the  trap.  The
can  is  designed  to contain  the flame  and  distribute  the heat
output  (e.g., about  100,000  Btu/hour) evenly  across the face of
the  trap.    Additionally,   the unit  provides  a  location  for
mounting  the  fuel  injection  nozzle,  ignition  plug  and  flame
sensor,  and  auxiliary  air   injection   nozzle.    Due  to  the
operating environment  and  required  long life, the burner can is
largely made of high  grade stainless steel.   The basic  cost of
this component  is  relatively  insensitive to  variations in heat
output  requirements.   Therefore,  the estimated  cost  of $21 is
used for all HDDEs.[7]

     The fuel  delivery system  is  composed  of a  fuel injector,
control  solenoid,  fuel  line,  and  fuel line connectors.   This
system  is  used  in conjunction  with the vehicle's existing fuel
injection  system.   The  use  of  an  electric  solenoid provides
precise  control   of  the  regeneration  rate,   and  provides
effective  overtemperature  protection.   The  fuel  injector  and
solenoid is  estimated at  $12  for  all  HDDEs.[7]   The fuel line
and  connectors  are  estimated  by  EPA  to  cost  about  $2  per

     The  ignition  system  provides  spark  ignition   and  flame
control.  The components  include  an electrode,  an inverter, and
a  step-up  voltage transformer  for  generating   a  high-voltage
discharge.   Also,  the system  includes a  flame sensor  and  sensor
relay  as  a  safety consideration  for cutting off  fuel  to the
burner  if  combustion fails.   Mueller Associates  estimated the
cost of  a  continuous  spark  system.[8]  The Agency  finds that a
somewhat smaller  system  providing  a  periodic spark,  when used
with the flame  sensor,  is fully satisfactory.  As a result,- EPA
estimates  the  system to  be  approximately 20 percent  less than
the'contractor's estimate, or $35 for all HDDEs.

     The  auxiliary air  system supplies a controlled  amount of
air  to the burner  and  trap during regeneration.   Its components
include  an air pump,  a  control  valve  operated  by  an electric


solenoid,  and  an air  delivery  line  with  connectors.  The  air
pump  is  criven  by  the vehicle  engine  and  is  a  larger,  more
durable  version of  those  already  used  on   light  trucks  and
heavy-duty  vehicles.   It  is  equipped  with  a  check valve  to
prevenc exhaust  backflow  into  the air pump.  Since  air  is  only
required  during regeneration,  the  air  pump  is  assumed to  be
equipped with  an electric clutch  so  that  it can  be disengaged
from  the  engine  in  order to  save fuel.  The  Agency estimates
the cost  of the air pump and  electric  clutch to be  about  $45
and the  associated  air delivery tubing with connectors at about
$5 for all  HDDEs.   The control valve  and  electric  solenoid  for
this system are estimated at $14 per vehicle.[7]

     The  exhaust  diversion  system   consists  primarily  of  a
bypass   valve   and    a   solenoid   controlled   actuator   that
temporarily   reroutes   the  exhaust   around   the  trap  during
regeneration.  The bypass valve is a butterfly type constructed
of stainless  steel,  located  just upstream of the combustor.   In
estimating  the  price  of  this  unit,  the  cost  of  a  stainless
steel  exhaust  pipe  has been included.  This  pipe  will replace
the standard  steel  exhaust  pipe that  normally  extends  from  the
engine manifold,  or turbocharger,  to  the  muffler.   In the DPS,
the cost of the stainless steel pipe  included  a  credit for  the
standard steel pipe which it  replaced.    In  this  analysis,  the
standard steel pipe  is  assumed  to  be roughly equivalent to that
required  to bypass  the trap.   Therefore, the  full cost of  a
stainless  steel exhaust  pipe   is  included in  the cost  of  .the
bypass valve.   This component  is  estimated at  $49  for LHDDEs,
$52   for  MHDDEs,   and  $58   for  HHDDEs.[8]    The  electric
solenoid/actuator is estimated  to  cost $15 for all HDDEs.[8]

     To  initiate  and   control  regeneration,  several  different
sensors, an electronic  control  unit,  and wiring harness will be
required.   A  backpressure  sensor  will  detect  the   need  for
regeneration  and is  estimated  to  cost $17 per  HDDE.[7]   A trap
temperature  sensor  estimated  to  cost $5,  and will  be used to
protect the trap from  excessive heat.[8]   A  sensor will also be
used  to  ensure  the engine  has reached the proper temperature
before regeneration  is  initiated.   This  sensor was estimated to
cost about  $1  in the DPS.

     Regarding  the  electronic  control  unit,  manufacturers  are
expected to equip  essentially  all HDDEs with  such  units  by the
1990s,  irrespective  of emission  standards.   For  this reason,
the  electronic capability required  for  trap  regeneration  will
be  added   to   the   existing  unit."  This   incremental  cost  is
estimated  at  about  $34  per  HDDE.[8]  The  wiring  harness  to
connect  the sensors  to the  electronic control unit  is estimated
to cost  $14  for  LHDDEs  and $18  for MHDDEs  and HDDEs.[8]


     Based on the above discussion,  the total cost of  the  fuel
eurner regeneration  syste.T  is estimated to be  S269  for LHDDEs,
$276 for  MHDDES,  and $282  for HHDDEs.   The  various  component
costs are summarized in  Table 3-33.

     Electric  Regeneration   System   Costs   —   This   type  of
regeneration system uses many  of  the same component parts  that
are  required by  the fuel  burner  regeneration  system.   These
include  an  auxiliary  air   supply  system,   exhaust  diversion
system,  and  electronic  control  system.   The exception  is,  of
course,  the  replacement  of  the  fuel burner  with  an electrical
heating  system.   The costs  for each  of  these systems  will  be
estimated below,  with the discussion  focusing on  the components
that  are different  from those  described for  the  fuel  burner

     The  auxiliary  air  supply uses  an air  pump and  electric
clutch,  as  required by  the  fuel burner system, except  that the
pump is  somewhat  smaller  in  size  because air  is  provided  only
to  the  trap.  The Agency estimates  this  to  cost about $40 per
HDDE.  The control valve and solenoid is  retained  at $14, as is
the air delivery line and connectors at $2 per vehicle.

     The  exhaust  diversion   and   electronic  control    systems
remain unchanged  from those  used  in conjunction  with  the  fuel
burner.   The exhaust diversion system  was estimated at $64 for
LHDDEs,  $67  for MHDDEs,  and  $73  for  HHDDEs.    The electronic
control  system was  estimated at $71 for LHDDEs and  $75 for the
other HDDE size categories.

     The  cost of  the electrical heating system is difficult to
estimate  because  the specific requirements  of the  system are
yet  to be well  defined.  The  Agency envisions a  rather modest
system   that  depends   on   the  vehicle's  existing  electrical
system.   In  this  concept,   the  additional  electrical   power
required  for regeneration  is  provided  by   using  the  existing
batteries  in  conjunction   with  a   larger   alternator.   The
electric  current  from the battery is  conducted by cable to the
electric  resistance  heating elements  which  are mounted  in the
trap  housing.   The  cable  is  equipped with  a fusable  link to
protect  the  batteries and charging equipment in  the event  of  a
short  circuit.    The  power   supply  to   the electric   heating
element  is   controlled  by  the  electronic control  unit  through
the  use  of  an  electromechanical  relay.   Based  on alternator
costs  supplied   by   Mueller  Associates,  EPA  estimates   the
incremental  cost  of  the  requisite  alternator to  be  about $47
for  LHDDEs,  $56 for MHDDEs,  and $67  for HHDDEs.  The cable with
a  fusable  link  is  estimated to  cost $7, while the  relay is
estimated  at $10  per vehicle. The  electrical  heating  elements
with mounting hardware  are  estimated to cost $14  for LHDDEs and
$18  for  the  other HDDE size categories.


                                              Table 3-33

                                HOPE Costs for Trap Regeneration System
Cost Category

Burner Can
Fuel Delivery System
Fuel Ignition System
Auxiliary Air System     ;
Lxhaust Diversion System '
Electronic Control System
Electrical System
Catalyst Dispensor System
Catalyst System Exhaust Mods
Fuel Burner
. 35
• -

$269    $287
$150    $166    $204



     EPA's estimated electrical heating system  costs  range from
S7S-S102,  depending  on HDDE  size.   Tr.e  Agency  believes  thec-3
costs are  representative of the type of electrical  systems that
are  commercially  viable  in  the  1990s.   However,   due  to  the
current lack  of  information  regarding  specific  designs,  these
cost   estimates   are  subject  to   a  significant   degree   of
uncertainty.     In   discussions   with  various  suppliers   of
electrical  heating  equipment,  Mueller  Associates  identified
several important uncertainties that  could significantly affect
the cost  of  the electrical system.   These  include  the  size  and
power  requirements  of  the  heating  elements, ensuring  adequate
battery capacity  for  engine  starting,  and prevention  of  trap
thermal stress  due to uneven  heating  during regeneration.   In
response,  Mueller  Associates  has  estimated the  potential cost
of  sophisticated  electrical  systems  that  address  each of  the
potential   areas  of   concern   as  suggested  by  its  industry
contacts.   Such electrical systems could  cost nearly four  times
more than that  estimated  by EPA.   The Agency believes that such
sophisticated systems  will be  found  to be  unnecessary  as more
information becomes available.  Hence,  only EPA's estimate will
be used in this analysis.

     As summarized in  Table  3-33, electric regeneration systems
are  estimated to  cost about  $269,  $287,  and  $306  for  LHDOEs,
MHDDEs, and HHDDEs, respectively.

     Catalytic  Regeneration  System  Costs  - Unlike  the  other
regeneration  systems,  which  are  at   least conceptually  well
known,   the  catalyst  and  the  technique which  will  be  used to
introduce   it  into   the  trap  remains   a   matter   of   some
conjecture.   The   approach assumed  in  this analysis  involves
onboard vehicle storage  of  a  metallic compound in  dry powder
form.   From   time  to  time,   a  metered  amount  of  catalyst  is
fluidized by  compressed air  from  the vehicle's turbocharger and
then injected into the exhaust  stream just  ahead  of  the trap to
initiate  combustion.   This  type  of  regeneration  system  is
potentially the simplest  with regard to  the required hardware.
The  primary  components consist of  the catalyst,  the  catalyst
dispensor  system,  and   the   electronic   control  system.   As
discussed  in  conjunction  with the  other regeneration systems,
the requisite stainless steel exhaust pipe  is treated as a part
of the catalyst regeneration system.

     The  metallic  catalyst is  assumed  to  be copper  in the form
of  powdered  copper  chloride  (CuCl) .   The  amount   of  catalyst
required  for  new  HDDEs is estimated to be about  2.3 pounds for
LHDDEs, 4.0 pounds for MHDDEs,  and  8.5  pounds  for HHDDEs.   This
is  based  on   the allowable maintenance  intervals  for each HDDE,
the  estimated fuel  economies  for  these  vehicles  as described
previously,  and an  assumed catalyst requirement equivalent to
0.16  g/gallon  of  diesel  fuel.[14]    Combining  this  with  an


 estimated cost  of  51.46/pound  of  CuCl,[8]  the  catalyst  costs
 are  estimated  to be  $5,  $9,  and  $13  for LHDDEs, MHDDEs,  and
 HHDDEs,  respectively.

      Tne catalyst  dispenser system  has several  parts.   A  one
 gallon  polyethylene bottle  is used  to  hold the  required  powered
 catalyst.   This is estimated to  cost  about $1 per  vehicle.[7]
 The  reservoir  will  be  attached to  a  metering  device  with  an
 integral agitator  extending  into   the  reservoir  to  break  up
 lumps  of catalytic  material.   This device is estimated  to  cost
 about  $30 for  each  HDDE.[7]  The metering  device uses  a  small
 electric motor with  an estimated cost  of  $12 per  vehicle.[7]
 After  being metered,  the catalyst  is  fluidized  and sprayed  or
 injected  into  the  exhaust  using  high   pressure  air.    The
 fluidizer/injector    unit   includes   the    required   fittings
 necessary to  mate  it  to the  metering device  and air  supply.
 The  estimated cost  is $7 per HDDE.[7]

      The final components  of   this  system  are   the  hoses  for
 transmitting  the  compressed  air and  fluidized  catalyst.   The
 Agency   estimates  the  cost  of   these   items  at  about  $3  per

      The  electronic  control  system   requires   four   sensors.
 Three of  these are  the same  as used  by  the   other  types  of
 regeneration systems:   trap temperature $5,  engine  temperature
 $1,  and  exhaust  backpressure $17.   An engine  speed sensor, is
 also used and has  an  estimated  cost of about $5  per  MODE. [8]
 The  electronic control  unit requirements  are again  incremental
 to the  existing  capability, although the cost is  somewhat  less
'than for  the  other  regeneration systems  because the  catalyst
 system   is   less  complex.   The  incremental electronic   control
 unit cost is  estimated by  EPA  to  be  about  $22  for all HDDEs.
 The  wiring  harness  will  also  be  less costly.   This   item  is
 estimated by EPA at  about  $9  for LHDDEs and  $12 for MHDDE  and

      The standard  steel exhaust  pipe   will  be   replaced with a
 stainless  steel  counterpart  as  with  the  other  regeneration
 systems.  However,  the catalyst regeneration system should  not
 require  that   the   exhaust   bypass   the   trap   during   the
 regeneration  process.    Therefore,  the  cost of the  stainless
 steel pipe  should  include  a credit  for   the deleted  standard
 steel pipe.   The incremental  exhaust   pipe  cost for each  HDDE
 size category  is taken from the corresponding estimate  used  in
 the  proposal,  with  one revision.  As described  in  the  DPS,  the
 incremental  exhaust pipe cost was  based on the  assumption .that
 about  25 percent  of  the  MHDDEs  and  HHDDEs would have  dual
 exhausts.  This assumption  has been  revised  because essentially
 all   MHDDEs   and  HHDDEs   are   expected  to  be  equipped  with
 turbochargers  in  the  1990's  and,   therefore,  will  likely  have


only one  exhaust pipe.   Making  this  revision,  the  estimated
incremental cost of the  stainless  steel exhaust pipe is 533 for
LHDDEs,  $42 for MHDDEs,  and $71 for HHDDEs.

     As  summarized in Table 3-33,  catalyst  regeneration systems
are  estimated  to cost  $150  for  LHDDEs,  $166  for  MHDDEs,  and
$204 for HHDDEs.

     Total Trap-Oxidizer Variable  Costs —  Table  3-34  presents
a  summary  of   the  three  trap  system  costs.    As  shown,  the
ceramic  monolith  trap with  a  fuel  burner   regeneration  system
has  an  estimated  cost of about $370  per  LHDDE,  $448 for  MHD.DE,
and $574 per HHDDE.   The ceramic monolith trap  with  an electric
regeneration system has  the potential of  costing about the same
as the  ceramic  monolith/fuel  burner  system.  The  ceramic fiber
trap with  a catalyst  regeneration system could  prove  to  be the
least expensive trap-oxidizer  with  estimated   costs  of  $223,
$272, and  $344  for  LHDDEs,  MHDDEs,  and HHDDEs,  respectively.
As   stated  previously,   the   ceramic  monolith/fuel   burner
trap-oxidizer  will   be  used  to  determine  the costs   of  the
particulate standard,  due to  the  greater  uncertainty associated
with the other designs.

     Now  that  the variable  costs for  each  HDDE  size category
have been  determined, the average  hardware cost  for  each trap
equipped  non-bus  HDDE and  urban bus,  as well  as that  for the
fleet-average  vehicle  can be calculated.   This   is  done  by
combining  information on sales  and  trap  usage  with the  system
cost  for   the   appropriate   HDDE   class   or   classes.   The
methodology for deriving  the  various average costs  is described
below.   This  methodology  will  be  used  in this   section  for
variable costs and in subsequent sections as  required.

     Identifying  the  variable cost  for  a  trap-equipped urban
bus  is  the most straight forward.   Due  to  their horsepower
ratings,  all  urban  bus  engines  generally can  be  classified as
MHDDEs.   In addition,  all urban bus  engines  will require a trap
oxidizer  to achieve   the  0.1  standard.   Therefore,   the variable
cost for  the  average  urban bus  is simply  the  value identified
for MHDDEs, or $448  (Table 3-34).

     The  average  cost for a  non-bus HDDE  with a trap oxidizer
is  found  by  sales  weighting  the  system  cost  for each  size
category.   As   described  elsewhere,  total  HDDE  sales  are
composed  of  35  percent  LHDDEs,  29   percent  MHDDEs,   and  36
percent HHDDEs.   However,  bus  sales  must be removed  from this
distribution  to find  the  percentage  of sales  in each category
for  non-bus  HDDEs  only.   Urban bus  sales  are  quite volatile
from  year  to  year,   but  would  generally  not   exceed  about   2
percent  of total  HDDE  sales.  Using  this  percentage  and the
fact  that  all urban  bus  engines are  MHDDEs,  the  non-bus HDDE


                        •Sable  3-J4

           HDDS Costs  for Trap-Oxidizer Systems

Ceramic Monolith/          Ceramic Monolith/
Fuel Burner Electrical Ceramic Fiber/Catalyst


sales account  for  98  percent of  the total  and  are distributed
as follows:  36 percent  for  LHDDEs,  27  percent  MHDDEs,  and  37
percent  HHDDEs.  Therefore,  sales  weighting  the  various system
costs with this distribution  results in a variable  cost  for the
average  affected non-bus  HDDE of $467.

     The per vehicle  cost when  averaged  over the  entire fleet
is the sum of  the  sales-weighted cost for an urban bus  and the
sales-weighted  cost  for   the  average non-bus  HDDE  (i.e.,  both
trapped  and  untrapped).   The urban  bus component  is simply  2
percent  of the cost  figure for a MHDDE.  The non-bus component
is 98  percent  of  the average  non-bus cost.   This  average, is
determined by  combining   the  various HDDE  category  costs  with
the  non-bus   sales  distribution  and  the  percentage  of  the
non-bus  fleet  that  is trap equipped.   Expressed  generically in
equation  form,  the   non-bus  component  of  the   fleet-average
vehicle  cost  is:

     Non-Bus  Portion  of Fleet-Average Cost »

(.98) X  [$  per LHDDE x  LHDDE Fraction)   +  ($  per MHDDE x  MHDDE
Fraction)  + ($  per  HHDDE x HHDDE Fraction)]  x (Trap Fraction)

     The last term  in this equation  adds  some complexity to the
calculation  since  the number  of  traps  required  for  non-bus
HDDEs is projected  to change  from about 70  percent in 1991,  to
about 60   percent  in  1994  if  the  0.25 were retained  in  that
year.  For ease of presentation,  the non-bus cost  component  of
the  fleet-average  vehicle  will be  evaluated as   a  short  term
value representing  70  percent  traps  and   a long  term  value
representing  60  percent  traps.   When  these non-bus components
are  combined with  the urban  bus component,  the  result  is the
cost  of  a fleet-average  vehicle in the short-term (1991) and
the  long-term  (1994).   The  fleet-average  cost  for  intervening
years can be  linearly  interpolated.   Using  this  methodology,
the  variable cost  of  the  1991 standards when averaged  over the
entire  HDDE  fleet  is  $329  in  the  short  term and  $284  in the
long term.

     iv.   Total Manufacturers  Cost  of  the 1991  Particulate

     The   total   undiscounted   and   discounted   costs   to
manufacturers  are  shown  in  Table  3-35.   The  fixed  costs are
reproduced  from  Table   3-28.   The  variable  costs  are  the
products  of  the hardware cost  per  fleet-average  vehicle and
HDDE sales as  shown  in Table 3-10.   The total undiscounted cost
is $420.9 million,  while the discounted cost  is $402.6 million.


                                     Tfcble 3-35

                              HDDE Manufacturers'  Cost
                         for the 1991 Particulate Standards

                                                  Undisoounted        Discounted*
"fear              RD&T          Variable Cost       Total              Total

1987             $8.01                  -              $8.01
1988             20. CM                  -              20.CM
1989             14.CM           -       -       -      14.CM
1990              7.94                  -               7.34
1991   .        .: -                  $126.CM          126.CM
1992              -                   125.31          125.3^
1993              -                   119.9M          119.94
     Discounted at  10 percent 10 1991.


      b.    Costs  to Users  for   HDDEs  Complying  with  the  1991
           Particulate Standards.

      i.    First Cost

      The  amount that  manufacturers  must  increase  the price  of
 each  HDDE depends principally on the variable cost per vehicle,
 the   number  of  vehicles  over  which  the  fixed  costs  will  be
 apportioned,  and on the  cost  of  capital  to  the  manufacturer.
 For  the 1991  standards,  it  is   assumed  that manufacturers  will
 generally recover their fixed costs  prior to the effective  date
 of  the more  stringent 1994  HDDE particulate  standard.   It. is
 further assumed  that the cost of capital is 10 percent.   Hence,
 the  first  cost  increase  for  a   vehicle  is the  sum of a  portion
 of  the  discounted  fixed  cost .and   the  hardware  cost,   as
 described earlier.

.  ._.  In  the   short  term  the   purchase   price   increment   for
 trap-equipped  non-bus  HDDE  is   estimated  at  $457  for  LHDDEs,
 $535  for MHDDEs, and $661 for HHDDEs.   This averages  $553 for a
 trap-equipped  non-bus  HDDE.   The  first  price  for  an urban  bus
 is  $535.   Expressed on a fleet-average  basis,  the  cost would be
 about  $390.    In  the  long  term,   assuming   a  manufacturer
 continues  to  charge  the same  per  vehicle for  its  fixed  cost
 recovery,  the  fleet-average  vehicle would cost  an  additional

      ii.   Fuel  Economy

      In  the  draft  economic  impact analysis, non-bus  HDDEs  with
 traps were estimated  to  incur   a  1  to  2  percent   fuel  penalty.
 Urban buses  were  assumed  tc  incur  a  2  percent  penalty.   As
 discussed  in  the Technical  Feasibility  Chapter, EPA's  estimates
 generally  fell  within the range  of  fuel penalties  that  were
 presented  in  the few comments on this  issue.  Also as  discussed
 in  that chapter, the  trap  volumes  in  this  final  analysis  have
 been   significantly   enlarged   from  those  assumed   in   the
 proposal.   As   a  result,  the   upper  range of  EPA's  previous
 estimate has  been revised downward  from a 2 percent  penalty to
 a -1.5  percent penalty;   Therefore,  the  new range for  non-bus
 HDDEs equipped  with  traps  is   1.0-1.5  percent,  while  the  new
 point estimate -for  urban  buses is  1.5  percent.

      Several  important methodological  changes have been  made in
 calculating  the fuel economy impact for each affected  vehicle.
 For  non-bus  HDDEs, the estimated  MPG  values for each  HDDE  size
 category  has  been  revised  to  reflect updated  estimates.   This
 was  fully  described in the previous discussion of  the 1991  HDDE
 NOx  standard where the discounted lifetime cost  of a  1  percent
 penalty for  each size category  was shown to be  the  following:
 $54    LHDHEs,    $259   MHDDEs,     and    $705   HHDDEs.     Using


 this   information,   the  1.0-1.5  percent  penalty  represents   a
 discounted  Lifetime cost of $35_0-$525   for  the average  non-ous
 HDDE  equipped  with  a trap.

      For  urban  buses,  three  key assumptions  have been  revised
 from  the  values used   in  the  draft  analysis.   A  recent EPA
 analysis  shows  the  average annual  bus  mileage is about  45,000
 miles rather  than  50,000  miles, and  that  the average  lifetime
 is  closer to  12  years  than 10 years. [3]  These two  changes are
 included  in  this updated  analysis.   The  third  change  affects
 fuel  costs  for  buses.   The  draft  analysis  utilized the  same
 cost  per  gallon of  diesel  fuel  for  both non-bus and  bus  HDDEs,
 i.e., $1.20/gallon.   Diesel  fuel  for  urban  buses  is  actually
 significantly  less  costly  than  that  for  non-bus  HDDEs  due  to
.volume   discounts   and   lower  taxes.   A  comment  from  the
 Department   of   Transportation   supported   the   use    of    a
 $1.00/gallon cost for urban buses.  This value has been  adopted
 for  use in  this  analysis.   Based on  these revised  values, the
.estimated  1.5  percent  fuel  penalty  results in  a  discounted
 lifetime  fuel  cost  of $1070 for  each urban bus.'  Expressed on  a
 fleetwide basis, the  average  HDDE  will incur  a  short-term fuel
 economy penalty  of  about  $261-$381 and  a  long-term  penalty  of
 about $227-$330.

      It should  be  noted that the  above fuel  economy penalties
 were  estimated for  trap-oxidizer systems using ceramic  monolith
 substrates   and  fuel burners  for  regeneration.   If  these  s.ame
 traps were  used with  an  electrical  regeneration,  the  penalty
 would likely  be about  the  same, due  to the  energy  required  by
 the  alternator.   However,   if  the ceramic  fiber trap  is   used  in
"the   future, " the fuel  economy penalty  would be somewhat  less.
 The  use of  a  catalyst  to  lower  the ignition  temperature  of the
 collected particulate  would  avoid the  use of energy intensive
 heating systems, since  the traps would   be  self  regenerating.
 In  this  case,  the   fuel economy  penalty would be  lowered  by  an
 amount  that  is  basically  equivalent   to  the  energy  used  to
 regenerate   the  other  two  trap  systems.   EPA estimates   this  is
 equivalent   to  about  a  0.5  percent  fuel   economy  penalty.
 Therefore,  the use  of  a ceramic fiber  trap might result  in only
 a 0.5-1.0   percent  penalty   rather  than  the  1.0-1.5   percent
 penalty used in this analysis.

      iii.  Maintenance

      The  draft  analysis  identified two  maintenance items  for
 trap-equipped HDDEs:   the  regeneration system  and  the  exhaust
 system.   The   regeneration  system  was   assumed  to   require
 maintenance after  approximately five  years  of  operation.   At
 that  point,  the engine temperature and trap  temperature sensor
 would need  replacement.  For  the  exhaust  system,  the customary
 replacement of  the  standard  steel exhaust pipe  was  expected  to


be eliminated,  because  this component  would  be displaced  by a
stainless  steel  exhaust  pipe  as   part  of  the  trap-oxidizer
system.   Sensor maintenance  when  discounted   to  the  year  of
vehicle  purchase  was  estimated  to  range  from $22   for  LHDDEs
with one  trap  to  $44  for HKDDEs with two traps.   A  net savings
was  projected  due  to  eliminating   the  need  for  exhaust  pipe
replacement.  This discounted  savings was  estimated   at  $39 for
LHDDEs, ranging up to $97 for HHDDEs.

     Four  general  comments  were   received  from  metropolitan
transit   authorities   and   the   Department  of  Transportation
suggesting  that maintenance  costs  would   increase  due  to .the
combined NOx  and  particulate standards.  Since the  maintenance
savings   that  were   projected   for   trap-based   particulate
standards  overwhelm  the  small  cost associated  with  the  NOx
standards, these comments would appear  to  be directed primarily
at the  former  standards.   The  New Jersey Bus  Operations,  Inc.,
specifically  claimed  that  EPA's  regulations  will  require the
use  of  electronic  control  units,  resulting  in  substantial
expenditures   for   training,    and   the   need   for   a   more
sophisticated  and  expensive labor  force.   In  another  specific
comment,   the  Department   of   Energy  estimated   that   the
particulate  standards   could  save  up   to  $402  for   a  HDDE. in
Classes  IIB-VI and up  to  $519  for  a  HDDE in  Classes  VII-VIII
(undiscounted).   Finally,   a  few  manufacturers  indicated  in
their   technical   feasibility   comments   that   a   trap   may
potentially   require  some   type  of   maintenance   during   the
vehicle's  lifetime  due  to  such  things as the accumulation of
ash or catalytic material.

     In response  to  the concern expressed  regarding  the forced
use   of  electronic  control   units  on   urban   bus  engines.
electronics  are projected  to  be  widely  used  on  all  types of
HDDEs   in   the   future   regardless   of   emission   control
requirements.   Hence,  costs  associated with  purported changes
in  the  labor  force  cannot   be  charged  against  the  emission
standards.   Concerning  DOE's  savings estimate and  the general
comments  that  maintenance  costs  will  rise,   EPA will address
these comments  by reevaluating  the  likely  effect on  maintenance
in. the  context of the  revised trap system design as described
in the  section on variable costs (i.e.,  ceramic  monolith/fuel
burner  regeneration system).

     The  assumption  in  the  draft  analysis   regarding  sensor
replacement  was  not  adversely  commented   upon,   and  is  being
generally  retained with a  few revisions.   The most   significant
changes  are  the use of new component  costs  and a  revision in
the   assumed   number   of   replacements  for   each   HDDE   size
category.  The  estimated retail  cost of each engine   temperature
sensor   is   $9 and   each   trap  temperature   is  $20. [8]   The
replacement  of both  sensors  is estimated  to  take one  hour at


 $28   per  hour.    To  be  consistent  with  the   revised   trap
 configuration,  all HDDEs  are assumed  to have a single  trap  ar.d,
 therefore,  only  one  trap  temperature  sensor  is  required  per
 affected vehicle.  To  account for the significantly  different
 lifetime mileages  of the  various  HDDE  classes,  the number  of
 replacements  over a  vehicles  life  has been  revised  to one  for
 LHDDEs,  two  for  MHDDEs,   and three   for HHDDEs.   Using  these
 values  and discounting the costs  at  10 percent over the  life  of
 the  vehicle,  the approximate  cost  is $35  for  LHDDEs, $57  for
 MHHDEs,  and  $71 for HHDDEs.

     A  review of  EPA's previously estimated  savings  for  exhaust
 pipes  has  resulted  in  this  category  being eliminated  to  be
 consistent  with  the  revised  ceramic  monolith  system  design.
 Again  referring  to  the  description  contained  in  the  variable
 cost  section, when the standard  steel  exhaust  pipe  is  replaced
 by its  stainless steel counterpart  (i.e., from the  manifold  to
^the bypass valve), the displaced  standard pipe is  assumed  to  be
'roughly equivalent to that  required  by the  bypass  system (i.e.,
 from  the  bypass   to  the   muffler).   For  the purposes  of  this
 analysis,  it  is  assumed  that  the  replacement  schedule of  this
 standard  pipe   will  be   roughly   equivalent  .regardless   of
 location.   Hence,  no incremental maintenance  is  estimated  for
 the exhaust  system.

     Regarding the  possibility  that  traps   could  require  some
 type   of  maintenance,  estimating  any  cost in  this  area' is
 especially difficult due  to  the  current limited information  on
 traps  themselves.  Nonetheless,  to cover the potential costs of
 such  a contingency,  EPA  will  assume  that trap maintenance  will
 cost  about $50 per  event  and that  frequency of  this  maintenance
 will  parallel  sensor maintenance.   Therefore,  when  discounted
 at 10  percent to  the year  of  purchase,  the cost  is $31  for
 LHDDEs,  $50  for MHDDEs,  and $62 for  HHDDEs.

     Total maintenance costs  for the  average ncn-bus  HDDE  with
 traps  is $102,  while the  cost for  an urban bus is  $107.   On a
 fleetwide  basis,  the cost per HDDE  in  the short-term  is  $72  and
 declines in  the long term to $62.

     As in  the  fuel economy  discussion,  it is appropriate  to
 examine the  potential maintenance costs  associated with  the two
 other   trap-oxidizer   systems.   The  ceramic  monolith/electrical
 regeneration   system   would   have    the    same    maintenance
 requirements and  costs  as the fuel, burner,  since  the same trap
 substrate  and sensors are used in both systems.  If  the ceramic
 fiber   trap-oxidizer  is   used in  the  future,  the  maintenance
 requirements would  be somewhat  different.   In  addition to  the
 costs   associated with  sensor and  potential trap  maintenance,
 catalytic  material  used  in  this  system may  need  replacement
 during'the vehicle's  lifetime.   Offsetting  these  costs would be


a credit resulting from  the  use of the stainless  steel  exhaust
pipe,  which  eliminates  the need  for  periodically  replacing the
standard  steel  pipe.   The   discounted   catalyst  replacement
costs,   as  estimated  using  the basic  methodology  presented  in
the variable cost section, are $14 for LHDDEs, $25  for  MHDDEs,
and $86  for HHDDEs.   The discounted  exhaust pipe  credits,  as
estimated  using  the  appropriate schedule  for  HDDV  standard
steel   pipe  replacements  described  in  the   DPS   are  $41  for
LHDDEs,  $51  for MHDDEs,  and  $81  for  HHDDEs.  Combining  these
values   with those  previously estimated  for  sensor  and  trap
maintenance  results   in  discounted  ceramic  trap  maintenance
costs  ranging  from $39  to  $138  depending on HDDE size.   .For
LHDDEs  this  is  significantly  less than the costs  estimated for
the ceramic  monolith/fuel  burner  system,  while  for the  largest
HDDEs  it is about the  same.

     iv.   Total User  Costs

     The total  cost to  the purchaser of an HDDE is composed of
the first  price increase and  the lifetime discounted  costs for
a   fuel  economy   and   maintenance.     The   cost   for   each
trap-equipped   non-bus   HDDE   is  $577-604   for   all   LHDDEs-,
$901-1,030 for  a MHDDE,  and  $1,499-1,852 for  a HHDDE.   For . the
average   non-bus  HDDE   with  • a  trap   the  total  cost  is
$1,050-1,180.  For an urban  bus it is $1,712.  Expressed  as an
average over the entire  fleet, the total user cost in the short
term is $723-843 and will decline  to  about  $625-728 in  the long
term.    The  fleetwide  cost per  vehicle is summarized  in Table

     6.     Total HDDE Manufacturer and User Costs  for  the 1991
           NGx and "articulate Standards

     The total  cost to  HDDE  manufacturers of  the  1991 standards
is  the  sum  of  the fixed  and  variable  costs of  the   NOx and
particulate  emission control.   These costs are passed on to the
users   of HDDVs  as  first  cost  increases,   and   are  added  to
operating  costs for  total user  cost of  the standards.  These
values   were  developed above,  and are presented in Tables 3-37
and 3-38.

     The  discounted  manufacturer cost  is about  $476  million,
while the  average increase  in  lifetime  user  cost for  a 1991
model year  HDDV is $803-1,171,  tapering off  to  $700-977 for  a
1993 model year HDDV.

     7.    1994  Diesel  Particulate  Standard  <0.10   g/BHP-hr

     In  this   section,   the   economic  effects   of  the   1994
particulate  standard  for  HDDEs  are  analyzed.   The costs are
examined  as  an  increment  to  those  that  would  result   from


                           Table 3-36

                         Toe a I  User  Cost
               for tr.e L99I Particulate Standards

            (Discounted  to Year of Vehicle Purchase)

                            	Fleet Average Vehicle
Cost Category               Short-Term               Long-Term

First Cost                    $  390                  $  336

Fuel Economy                   261-381                 227-330

Maintenance                      72                      62

Total                         $723-843                $625-728


                          Table 3-37

                 Total HDDE  Manufacturer Costs
              1991 NOx and  Particulate  Standards

Und is counted
Grand Total
$28. 7M
49. 5M

$42. 3M
371. 4M
413. 7M
$491. 9M
$34. 8M
63. 4M
98. 2M

$38. 5M
339. 1M
377. 6M
$475. 9M
*    Discounted at 10 percent to 1991
**   Model year 1991-93 HDDVs.

                              3-92  -

                           Table  3-38

                     Total HDDE User Cost*
               1991 NOx and Particualte  Standards
NOx    _ .



Grand Total


Short Term
Long Term
Fuel Maint-


62 -

     Incremental cost over cost of 1988 standards.


continuing  the  1991  standard  into  the  1994  and   later  model
years.  As  discussed  in the  Technical  Feasibility  Chapter,  tne
Agency  expects  that  the  1991  0.25  g/BHP-hr  standard  would
require  about  60  percent  of  the  non-bus  HDDEs   to  be  trap
equipped in  the long  term  (i.e.,  about 1994).   Similarily,  the
1991  0.10 g/BHP-hr standard would require all urban bus engines
to  be  trap  equipped.    With  the  more  stringent  1994  0.10
g/3HP-hr,   about  90   percent  of  the  non-bus  HDDEs will  need
traps,  while   the   buses  will   all   remain   trap  equipped.
Therefore,  the  incremental  effect  and  resulting  cost  of  the
1994  standard  is  dependent  on  the  use  of  an  additional  30
percent traps by non-bus HDDEs.

     The basic  inputs for  this  analysis are taken from  the 1991
particulate  standards  section   where  the  costs   of  various
trap-oxidizer   systems   were  examined.     Specifically,   that
analysis reviewed the costs  associated with  a ceramic  monolith
trap  using   a  fuel  burner   regeneration  system,  a  ceramic
monolith  trap using  an electrical  regeneration system,  and  a
ceramic  fiber trap using  a   catalyst  regeneration  system.   The
details of  that comprehensive  evaluation will not  be  repeated
here.   It  is important  to note,  however,  that only the ceramic
monolith/fuel burner  system was  used  to  estimate  the  econgmic
effects  of   the  1991 standards.   This approach was  taken,  in
spite   of   the   potentially   lower   cost   of   the   ceramic
fiber/catalyst system,  because  the  ceramic monolith/fuel burner
trap  is  presently the  most  well  defined  and may be  the first
commercially available  trap-oxidizer.

      The  economic effects  of  the  1994  standard will  also be
assessed   using  the   costs   associated   with  the   ceramic
monolith/fuel   burner   system.     Due   to   the  long   leadtime
associated  with  the  1994  requirement,  however,  it  is  possible
that   a  lower   cost   trap  oxidizer   such   as  the  ceramic
fiber/catalyst  system may  be  widely used by  the effective date
of  the  standard.   If this  were  to occur,  the cost  of  the  1994
standard  would  be  somewhat  less  than  that  presented  in  the
subsequent  sections.                 -  -

      a.     Cost to the  Manufacturers

      i.     Fixed Costs

      The  total  fixed cost of the 1994 standard obviously will
be  significantly  less  than  that  associated   with   the   1991
standards.   Only  30  percent  of  the HDDEs will incur development
and  certification  testing costs,  compared to  about  70 percent
in  1991.    Also,  when  reviewed  on a  per  vehicle  basis,  it is
reasonable   to   expect   that  engineering   experience  gained
throughout  the  early 1990's  will make the application  of traps
to  new  families in  1994  less  difficult  than it was  in 1991.


 Therefore,  the fixed cost that was  recovered  in  the  sales  price
 of  a trap-equipped  HDDE  under .the  19S1  standards (i.e.,  about
 $87)  would seem  to  represent  an  upper limit for  the  fixed  cost
 associated  with the  1994  standards.

      Using  this  conservative  assumption,  the  total  fixed  cost
 of  the 1994  standards  can be  estimated  by multiplying  $87  per
 trap-equipped vehicle by  the  number of such vehicles  over  which
 fixed costs  are  recovered.   As  in  the previous  HDDE  analyses,
 the  fixed costs  of  a standard  are  assumed to be  recovered  over
 three years  of  production  immediately following the  effective
 date of  the  standard.   The  estimated cost  of   capital  is  10
 percent  per  annum.   Based  on the projected HDDE  sales  in  Table
 3-11,  the number of  trap-equipped  HDDEs  used  in the  fixed  cost
 calculation  is  about  334,000  (i.e.,  30  percent of  the  1994
 through  1996 HDDE  "discounted  sales," excluding buses).   This
 results  in total estimated  fixed  costs  of  $29.1 mi 11 ion, """when
 expressed as  a lump sum investment  in 1994.   The  expenditures
-of  these over  time  can  be  expected to occur as  shown  in  Table

      ii.    Variable  Costs

      The  variable cost of a  specific trap-oxidizer  system  is  a
 function  of vehicle  size.  The costs of  a ceramic monolith trap
 with a  fuel burner  regeneration  system was  previously estimated
 at   $370  for  LHDDEs,  $448   for  MHDDEs,   and   $574  for  HHDDEs.
 Total HDDE  sales,   excluding  urban  buses,  are  composed of  36
 percent   LHDDEs,   27  percent MHDDEs,  and  37  percent  HHDDEs.
 Using these  values  to  sales weight  the  trap-oxidizer  cost  for
 each size  category  results  in an average variable cost  of $4-67
 per   trap-equipped   HDDE.   Expressed  as  an  average  over  the
 entire  fleet, the variable  cost is  $137 per HDDE.

      iii.  Total  Manufacturers   Cost  of  the  1994   Particulate

      The    total    discounted    and    undiscounted    costs   to
 manufacturers  are   shown  in Table  3-40.   The  fixed  costs  are
 taken directly from Table  3-39.  The variable  costs  are  the
 products  of  the  hardware  cost  per  fleet-average  vehicle  and
 total HDDE sales in each specific year (Table  3-11).   The  total
 undiscounted cost is $193.5  million,  while the  discounted cost
 is  $217.9 million.

      b.     Cost  to   Users  for HDDEs  Complying  with  the  1994
            Particulate Standards

      i.     First Cost

      The  amount  that a manufacturer must  increase  the price of
 an HDDE  to recover  its  expenses depends on  the timing of  the
 costs,  the cost of capital,  and  the number  of vehicles  over

                           Taole 3-39


                    Torrai Fixed Cose  of  the
                 1994  HDDE Parciculate Standard
Fixed Costsfl]
$ 5.0 M
13.0 M
7.9 M
3.2 M
Fixed Costs
$3.4 M
9.8 M
6.5 M
2.9 M
$29.1 M
$22.6 M
"[Tl  Discounted  to  the  effective date  of  the  standard,
                                i .e.,

                             Tacle  3-40

                      HDDE  .".anu£-3Cturers '  Cose
                 for the !?">•> Particulate Standard

                                       Undiscounted     Discounted
Year    Fixed Cost     Variable  Cost        Total	       Total

1990      $3.4 M             —           $  3.4 M        $  5.0 M
1991       9.8 M             —              9.3 M          13.0 M
1992       6.5 M             —              6.5 M           7.9 M
1993       2.9M             —              2.9M           3.2M
1994        —           $56.0  M          56.0 M          56.0 M
1995        —             57.0  M          57.0"M          51.8 M
1996        —             57.9  M          57.9 M          47.8 M

Total  •                                  $193.5 M      ..  $184.7 M


which  the  fixed  costs  will   be  recovered.   As  discussed  in
deriving the fixed costs, rranufacturers  are expected to recover
fixed costs over the  first three years of  production.   The cost
of capital was  also  identified as 10 percent per annum.  Hence,
the first cost  increase  for  a  vehicle is  the  sum of  a portion
of  the  discounted  fixed  cost   and  the  hardware  cost,  as
described earlier.  Using this methodology, the  purchase  price
increment  for   a  trap-equipped HDDE  is  estimated  at  $457 for
LHDDEs, $535 for  MHDDEs, and  $561  for  HHDDEs.   This  averages
$553  per  trap-equipped MODE.  Expressed on a fleetwide  basis,
the purchase price increase  is  about $163.

     ii.   Fuel Economy

     Traps may adversely affect fuel  economy due  to  a potential
increase  in exhaust   backpressure  and  because  of   the  energy
required to initiate regeneration.  The  penalty  associated with
the   use   of   this  technology   was   estimated  in   the   1991
particulate  standards   discussion   as-  1.0-1.5  percent  per
trap-equipped vehicle.   When  discounted to  the  year of vehicle
purchase, this  is equal  to  approximately  $54-81 for  a  LHDDE,
$259-388 for a  MHDDE,  and $705-1,058 for a HHDDE.  This amounts
to  $350-525  for  the  average  trap-equipped  HDDE.   For .the
fleet-average HDDE, the  discount  lifetime  fuel penalty is  about

     iii.  Maintenance

     The  potential  maintenance  costs   associated   with  trap
oxidizers   fall   primarily    into    two  categories:     sensor
replacement  and  trap  maintenance.   The  discounted  lifetime
costs  associated  with  these  items were estimated in the section
on the  1991 particulate  standards as being $66,  $107,  and $133
for  a  LHDDE,  MHDDE,  and  HHDDE,  respectively.   For  the average
trap-equipped  HDDE this  is   $102.    Expressed on   a   fleetwide
basis,   the   discounted  lifetime   maintenance   increment  is
estimated $30.

     iv.   Total User Costs

     The total  cost  to the  purchaser of an  HDDE  is  composed of
the  first price increase, and  the discounted  lifetime  costs for
fuel  economy   and  maintenance.    The   total   user   costs  for
trap-equipped   HDDEs   complying   with  the  1994  standard • are
$120-147,  $366-495,   and $838-1,119   for  HHDDEs,  LHDDEs,  and
MHDDEs,  respectively.   For   the  average HDDE  with  a trap, the
total  cost  if   $1,005-1,181.    Expressed  as an average  over the
entire  fleet,   the  total user cost  is  $296-347.   The various
fleetwide costs per vehicle  are summarized in Table  3-41.

                   Taoie 3-41

             Total  '-'sec-Cose  for  ens
            1994  Particulate  Standard

    (Discounted co Year of Vehicle Purchase)

Cost Category                Fleet-Average Vehicle

First Cost                           $163
Fuel Economy                        103-154
Maintenance                           30

TOTAL                              $296-347


     8.    Aggregate Costs  to the  Nation of  the  HDE NOx  and
           Particulate  Standards

     The  aggregate  costs  to  the   nation of  the  HDE NOx  and
particulate standards  include  the   total  manufacturer  costs  of
RD&T   and   hardware,   and  user   costs   of   fuel   economy  and
maintenance which  will  be  incurred  due to  the  more  strict
emission  control  requirements of   the  standards.  These  costs
were developed  above,  and  are shown  in Tables  3-42  and  3-43
according  to  the  year of  expenditure.  All  costs  before  the
year of the standard are for RD&T,  including  certification,  and
costs  after  the  year  of  the standard  are  for  hardware  and
additional operating costs  for the  venicles equipped  with HDEs
projected to be sold in those years.

     The  aggregate  costs  presented in  Tables  3-42  and 3-43  for
each model year group are incremental  in  nature.   The aggregate
incremental costs  for  the 1991 model  year  group  represent only
the  added  costs  beyond those  incurred in the 1988  model  year
group.    The same is  true  in  considering the 1994  model  year
group aggregate costs,  with the exception that the increment is
calculated relative to  1991.

     All  costs  are   shown   undiscounted  in  Table  3-42  and
discounted at 10 percent  to the  year  of  the  standard in Table
3-43 and are developed  in  the  preceding sections.   As  shown,
the  aggregate  costs  to   the  nation  of  the  HDE   NOx  and
particulate  standards  are  approximately  $118-600 million  for
the  1988  standards,  $833-1,241 million for  the  1991 standard,
and  $336-394  million  for   the   1994  particulate   standards,
discounted to each of those years.


     D.    Socioeconomic Impacts

     The  socioeconomic   impact  section   in   the  Draft   RIA
discussed the  effects  on manufacturer sales and  cash  flow,  the
regional  effects  of  employment,  and  the  national  effects  on
vehicle   purchasers,   energy  usage,  balance  of  trade,   and
inflation.   These  effects will not change significantly as  a
result  of   the   reanalysis   of  costs,   since  cost  estimates
decreased or rose  only  slightly from  the original  estimates.
However,   some   comments    were    received    from   citizens,
environmental groups,  the American  Trucking Association (ATA),
and    public   transit   system   operators    concerning   the
socioeconomic '  impact   of    costs    on    individuals    and
organizations.   The  questions  raised   by these  comments  are
reviewed  in  the following paragraphs.

     Comments  received from  citizens  and  environmental groups
argued  that  the  cost  of  these  regulations are rightly passed on
to  the consumers  who  also   receive the  benefits of   improved

                           Table 3-42

               Undisccunted Aggregate Increrer.tal
                   Costs 3t the HDE Standards

                      (millions  of  dollars)
Model Year 1988

HDDE NOx HDDE Particulate
Model Year 1991
Model Yea

HDDE NOx HDDE Particulate
r 1994

HDE Particulate

                                           HDGE NOx

                                           HDGE NOx
HDE Total

HDE Total


Model Year 1988
                            Table 3-43

                 Discounted* Aggregate Encrerrental
                    Costs oc the HDE  Standards
                       (millions of  dollars)
                                           HDGE NOx
Model Year
Model Year

HDDE NOx HDDE Particulate
- 14.0
9.7- 88.8


HDDE NOx HDDE Particulate
11.9- 69.8


Year HDE Particulate


 HDE Total

  16.1- 95.2
                                            HDGE  NOx
 HDE Total

      10 percent  to  r.r\e  year  of  the  standard.


 environment  and public health, and  thus  "will  think that it  is
 worth  the cost."  The  Agency agrees with  these  corrjnents.  EPA
 expects  the manufacturers to  recoup their losses through first
 price  increases  in LDTs and HDEs.

     On  the opposite side  of the argument,  ATA  believed  that
 the  costs  may  be  too  high,  stating  that,   "at  issue  is not
 whether  the average  motor carrier will be  adversely impact but
 rather,   in  the  case  of  fuel  penalties  for  example,  the
 magnitude of  this  effect  at the upper  end of  the" range  in
 potential penalties  on  the  highest mileage  group  of   single
 truck  or  small  fleet  owner/operators."   In response, the  Agency
 believes  that  the ATA has posed  an  unrealistic scenario.  There
 is no reason  to expect  the  maximum operating cost  impact  to
 fall   on  small  high-mileage  operators,  since  these  operators
"will  certainly search the market  for the  vehicles with  minimal
 fuel  economy impact  if  operating cost  is  of great concern.   A
 comment  by the  National  Resources Defense Council  is relevant
 here,  which  states   that,  "even  the more  expensive  standards
 still  add  only  a  small  fraction   to   the  initial   cost  and
 lifetime  operating cost  of  the  vehicles  in question."  Costs
 should be able  to be  easily  borne by the trucking industry  with
 small  increases  in  the  prices  of  consumer  goods;  since  these
 costs  will be carried by  all  segments  of  the  industry,   no one
 group  should receive an unfair advantage  or  disadvantage due  to
 the  standards.

      Several  comments were  received pertaining specifically  to
. the  socioeconomic impacts  of the proposed NOx and  particulate
 standards   on   urban   transit  Anises.    The    Urban   Mass
 Transportation  Administration  (UMTA)   and  local  transit   and
 transportation   authorities   from   New   Jersey,   Washington,
 Chicago,  Cleveland,   Washington,   D.C.,   Albany,  and San  Antonio
 all  stressed  the economic  burden  that  would be  placed  upon
 urban transportation locally  and  nationally.   There was  general
 agreement among these agencies that  EPA underestimated  the  true
 costs  and  economic   burdens  associated   with   the   proposed
 standards.   New Jersey Transit and  VIA  Metropolitan Transit  in
 San   Antonio  indicated  that  the   increased   costs  would  be
 translated into  higher* fares, lower  riderships,  more  personal
 vehicle   use,  and an  increase in emissions  as  a net  result  of
 the  proposed standards.  Finally, the Chicago  Transit  Authority
 (CTA)  expressed concern that  engine  selection  for transit buses
 would be reduced as  manufacturers leave  the market due  to the
 increased costs of control.

      EPA has estimated the  first  price  increase associated with
 a 0.10   g/BHP-hr  particulate and 5.0  g/EHP-hr  standard  at  a
 value of  $644.   The total  fuel  economy penalty  resulting  from
 these controls  is estimated  to  be 2 percent,  or $1427,  in the
 long  run,  and  slightly  higher in the short  run.  There  is  also


a  rsaintenance  cost  cf  $107   per   bus   associated  with  the
particulate  standard.   Wich  the  current  average  price  of  a
diesel  transit  bus  being   $135,000-145,000,  the  first  price
increase estimated  represents  at most a 0.5 percent increase in
the  first  price  of  a  diesel  transit  bus.   The  operating  and
maintenance cost  associated  with an  urban transit bus will rise
at most slightly over 2 percent.  This  assumes  that  fuel is  the
only operating  cost involved;  other  considerations would reduce
this figure.   Thus,  the  "economic  burden"  associated  with  the
NOx  and particulate  standards  does  not  appear  to EPA  to  be
severe.  Based on this, EPA does not  believe  that there will be
any  significant  fare increases  and  associated ridership losses
attributable to the standards.

     The market for  diesel  engines  used in  transit  buses  is
small  as  CTA  has   indicated.   Currently  only  one  domestic
manufacturer, GM, makes engines  for   large urban  transit buses,
and  only  one  or  two  of  their  five  such  engines  are  made
expressly  for  that  purpose.   Also,  Caterpillar makes  an engine
used in  smaller transit  buses  in some urban  areas.  EPA feels
that  it  is  highly  unlikely  that   either  manufacturer  would
relinquish its share  of the market under such circumstances.



     1.    "Cost  Estimations   for   Emission  Control   Related
Components/Systems and Cost  Methodology Description,"  Rath  and
Strong Inc.,  EPA-460/3-78-002,  March 1978.

     2.    "Diesel  Particulate  Study,"  U.S.   EPA,   OAR,  QMS,

     3.    "Heavy-Duty  Vehicle   Emission   Conversion   Factors
1962-1997," Mahlon C.  Smith IV, EPA-AA-SDSB-84-1,  August 1984.

     4.    "Standards  for  Emission  of  Particulate Matter  From
Diesel-Powered Light-Duty Vehicles  and  Light-Duty Trucks; Final
Rule" (49 FR 14496,  March 5,  1980).

     5.    "Standards  for  Emission  of  Particulate Matter  From
Diesel-Powered  Light-Duty  Vehicles  and  Light-Duty  Trucks  and
Technical  Amendment  to   Emission  Regulations  for  Light-Duty
Vehicles,  Light-Duty   Trucks,   and  Heavy-Duty  Engines;  Final
Rule" (49 FR 3010, January 24,  1984).

     6.    Oral  Communication  with  Terry  Ullman,   Southwest
Research Institute,  San Antonio, Texas,  January 28, 1985.

     7.    "Cost  of  Selected   Trap-Oxidizer System  Components
for  Heavy-Duty  Vehicles,"  Jack Faucett  Associates and Mueller
Associates, September  28, 1984.

     8.    "Costs of Selected  Heavy-Duty  Diesel  Engine Emission
Control  Components,"  Jack   Faucett  Associates   and  Mueller
Associates, February 8,  1985=

     9.    "1983  and  Later   Model  Year   Heavy-Duty  Engines,
Proposed Gaseous  Emission  Regulations:  Summary and Analysis of
Comments to the NPRM," EPA, OMS, ECTD, December 1979.

     10.   "Cost  Estimations   for   Emission  Control  Related
Components/Systems and Cost  Methodology Description,  Heavy-Duty
Trucks," Rath & Strong,- Inc., EPA-460/3-80-001, February  1980.

     11.   "Light-Duty   Diesel   Organic  Particulate   Control
Technology   Investigation,"    Southwest   Research   Institute,
EPA-460/3-82-011, August  1983.

     12.   "Control of  Air  Pollution  From  New  Motor Vehicles
and  New  Motor  Vehicle   Engines;  Federal  Certification  Test
Results For 1984 Model Year," U.S. EPA,  OMS,  CD.

Diesel  Bus
Thomas M. Baines,
Preliminary  Particulate
Engine,"  Terry  L. -Ullnan,  Charles
      SAE No. 840079,  February 27,
Trap  Tests  on
     14.   Memo  from  Charles  L
Request  Regarding a  Manganese
Trap Regeneration System, U.S.
                   .  Gray,  Jr.  to Robert
                    Fuel Additive-Based
                   EPA,  QMS.
a  2-Stroke
 Hare,   and
                Maxwell,  VW

                           CHAPTER 4


     This chapter will  examine the environmental  effects which
can  b.e  expected  to  result  from the  implementation   of  the
revised  NOx  standards  for  light-duty  trucks  and  heavy-duty
engines  and  new diesel  particulate  standards  for  heavy-duty
diesel  engines.   The material presented  here  begins  with  an
overview  of  Chapters 4  and 5 in the  Draft  Regulatory  Impact
Analysis, followed  by a  summary  and  analysis  of  the  comments
made  on  the  information  contained   in  these  chapters,  and,
finally,  a   presentation   of  revised   projections  of   the
environmental  and air  quality impacts  of  the  NOx  and  diesel
particulate emissions.

I.   Overview of NPRM Analyses

     A.    Oxides of Nitrogen  (NOx)

     The  Draft  analysis  opened  with  a   brief  review   of  the
health  effects  associated  with   NOx emissions.   The   primary
concerns  reviewed  were  the  human   respiratory  effects  which
formed  the  basis for  the  level of the  primary ambient  N0'2
standard.   At  the  present  time,  this  standard   level  is  an
annual arithmetic mean of 0.053 ppm.

     Following  this  review,  the effect  of  the   proposed  NOx
standards on ambient  air quality  was  estimated   by comparing
future  year  NOx  emissions   inventories  and ambient  NO2  levels
under  three  scenarios:   1)  no future control,  2)  the proposed
standards,  and  3)  the eventual  standards  as mandated  in  the
Clean  Air Act.   These  analyses  focused  on  those urban areas
that  are  wichin range of  exceeding the NAAQS  by the  end of the
century.    In   addition,    estimates  of   lifetime  emission
reductions per  vehicle  were  made, primarily for use  in the cost
effectiveness analysis.

     The  air quality analyses  for NOx were performed  using   a
three-step   approach.   The  first  step  involved  the   use  of
MOBILE2.5  to estimate  emission  factors  by  calendar year  and
vehicle  class  under the  three scenarios.   MOBILE2.5 determines
emission  factors in grams per mile   (g/mi)  for  motor vehicles,
based  upon  vehicle class, engine type,  model year,  and age of
the  vehicle.   For  heavy-duty  engines,  additional factors  are
used   to  convert  brake-specific  emission  factors  to  g/mi
emission  factors.   In order  to obtain a specific  calendar  year
emission   factor    for    the    individual   vehicle   classes,
dieselization  rates  by model  year,   registration  distributions
by  age,  and  mileage  accumulation  rates by  age are combined with
the  emission factor  by model  year   and  age.   The  model  year


 emission   factors   reflect  inproverrents  in  control   efficiency
 over   time.    The   calendar  year  emission  factors  are   then
 utilized  by  the  EPA Rollback  Model  in  step  3, described  below.

      In  the  second step, base year  inventories  of  NOx emissions
 for  the urban areas of  interest  were obtained from the  National
 Emissions Data System (NEDS).[1]  NEDS provides  county-specific
 estimates of  emissions  by  source  category for  each  county  in
 the   United   States.   Total  vehicle  miles travelled  (VMT)  by
 county,  VMT  breakdown  by vehicle class  by county, and  vehicle
 emission   factors   are  the  key  parameters in  determining  the
 mobile source inventory.  The 1981  NEDS  inventory contained, in
 the   draft   analysis  was  derived  using  emission  factors  from

      These were  combined,  along  with current  NOZ  levels  and
 projected growth in source activities  and  control  efficiencies,
 to yield future  year  emissions  and  NO*   levels.   This  final
.step was  performed using  the  EPA  Rollback Model which  begins
 with  base  year  inventories  of  NOx  emissions' and  base  year
 ambient    levels   of    NOZ  concentration   (design    values).
 Utilizing the emission  factors  from  the MOBILE program,  along
 with  projections  of  total  VMT  by  vehicle type,  and  similar
 numbers   for  stationary  sources,  the  model  can  then   project
 future year  inventories  of  NOx  emissions  and  corresponding
 ambient   levels  of  NOZ.    The   emissions  from  the   various
 sources  are discounted to  reflect  their  impact  upon  air qual.ity
 in the immediate  local area.   Increases  in ambient  NO?  levels
 are  assumed  to  move linearly with  increases  in discounted  NOx
Remissions.                                  ,

      Estimates  of   lifetime  reductions  in  NOx  emissions  per
 vehicle    were   calculated   in   a   straightforward   manner.
 Differences  in the emission  factors  by mileage  for the  various
 control   scenarios  and  estimates of  mileage  accumulation  over
 time  for  the   appropriate   vehicle  classes   (obtained   from
 MOBILE2.5)  were combined and  summed over  the vehicle's life.

      A more  complete  description of  the  modelling  procedures
 can  be found  in  the  Draft  RIA,  and in the  following  documents:
 "User's   Guide  to  MOBILE2",[2]   "Compilation  of  Air Pollution
 Emission   Factors:   Highway  Mobile Sources",[3]  and  "Rollback
 Modelling:   Basic and Modified".[4]

      B.     Particulate  Matter

      The  Particulate Environmental  Impact  Chapter in the  Draft
 RIA   opened  with  a  discussion  of  the  relationship  of  diesel
 particulate matter to  total  suspended  particulate  and the NAAQS
 for   particulate  matter.   The widespread  non-attainment of  the
 NAAQS in 1995,  under  either the  current  TSP  standard or  the
 proposed PM10 standard, was emphasized.

      Following  this  discussion,  the  lifetime  reductions  in
 particulace emissions  per  vehicle were  then  derived,  a'-ain for
 use in the  cost-effectiveness  analysis.  These  reductions were
 estims-ted  using  the  same  basic  methodology  as  that  described
 above for the NOx analysis.

      Next,  nationwide  and  nationwide-urban emissions  of diesel
 particulate were  presented.   These  projections  were made using
 the  same  basic  methodology  as  for   NOx,   but  with  slight
 modifications.   For instance, due  to  the  widespread violation
 of  the  particulate NAAQS,  it  is not  reasonable to  model each
 urban  area  individually.    Thus,  all  U.S.  urban  areas  were
 analyzed  together.   Also,  the  MOBILE  model  itself  is  not
 equipped  to determine  emission  factors  for  diesel  particulate
 matter,  so it  could  not  be  used  in  the  diesel  particulate
 analysis.   However,  the concepts  of MOBILE  and all applicable
 parameters  contained  in MOBILE2.5 (described  in detail  in the
 Diesel  Particulate  Study,  or   DPS[5])  were  used  to  estimate
 calendar year emissions.

      Since  the  diesel  particulate  analysis   is   done  on  a
 nationwide, and  not  on an  individualized urban  area basis, NEDS
 is  not   used  as  the  source  for  the  base  year   inventories.
 Instead,  emission  factors  were  combined  with  estimates  of
 nationwide  urban  VMT  by  vehicle class  to  develop  base year
 inventories of diesel particulate emissions.

      The  estimates of  nationwide emissions  were  then followed
 by  projections  of  ambient diesel particulate   levels.   Due to
- the  difficulties  in  distinguishing  diesel  particulate  from
 others  in  atmospheric  measurements, some  measurable surrogate
 in  ths  ambient  air  that  is  directly  relatable to  vehicular
 emissions  must  be  used  to  estimate  current  ambient   diesel
 particulate levels.   The  two surrogates  that have  historically
 been used  are  lead and CO.  Three types  of  ambient  impact were
 addressed:    1)   levels   expected   to   occur  at   air  quality
 monitors,  2)  average  exposure levels of urban  dwellers,  and 3)
 ambient  levels in  selected high-exposure situations.
 *                   >.
      In  estimating  urban monitor-type  levels,  conceptually,
 historic  ambient  lead  levels are first converted  to historic
 ambient  diesel  particulate  levels.   This  is done  by assuming
 that the ratio  of  ambient  concentrations of  the two pollutants
 is  equal to the  ratio of  their  emissions,  taking  into account
 that a certain fraction  of leaded particulate emitted falls out
 of  the atmosphere very  quickly  and  does  not  affect  ambient air
 quality.   Future  ambient  diesel  particulate  levels   are then
 projected   from  historic  levels  using  the  general  rollback
 approach.   Projections were made  for  a broad  spectrum of city
 sizes  and meteorological conditions.


       Annual  average urban exposures, which include  a  vanecy of
   individual  activity  pattern  elfects,  were  based  on  a  model
   developed  by EPA  to  estimate exposures under various  levels of
   the  CO  NAAQS.   The  model  was  based  on  measured  exposures in
   specific types of  situations in four  U.S. cities,  and  involved
   placing  the  population into various cohorts which spend  various
   amounts  of  time  in   each  exposure situation.   The  CO  levels
   projected  by  the  model  were  converted   to  diesel  particulate
   analogously   to   the   conversion   described   above   for  the
   lead-surrogate  model.

      »The high-exposure,  or microscale, situations were  analyzed
   using  models   developed  for  EPA   for  the  projection  of any
   completely  dispersed,  non-reactive pollutant.   Thus,  they are
   also   based  on  the   surrogate  and  rollback  concepts.    Four
 "situations  were modelled:   roadway  tunnels,  street  canyons, on
   an expressway,  and nearby an expressway.

       Following     these     three    estimates   "of    microscale
   concentrations  of  diesel particulate,  the particular  need to
   control   diesel particulate  at  high  altitude  was  discussed.
   While  the lack of particulate  emission  data  at high  altitude
   prevented any  more  precise  estimate of  environmental  impact
   than that presented in the nationwide  analysis  described  above,
 -  Denver's air  quality situation  was discussed  briefly and the
   need for high-altitude control was  established.

       Following  these  emission  and air  quality  analyses, the
   Draft  RIA attempted  to put these projections in perspective by
—examining four  classes of health and welfare  effects  associated
'"•'with   diesel particulate:    1)  non-cancer  health  effects, 2)
   carcinogenic health effects, 3) visibility,  and  4) soiling.

       The analysis of  non-cancer  health effects  associated with
   diesel particulate focused  on  identifying  the potency of diesel
   particulate  relative  to  that  of  general  suspended   inhalable
   particulate  (i.e.,   PMi0).   Using  this   relative  potency, the
   ambient   diesel  particulate   levels   identified  earlier   were
   compared to  the  current  PM10  levels  of urban  areas and the
   proposed PM10  standards.

       With respect  to  carcinogenic  effects, an  estimate of the
   lifetime  risk  of  contracting  lung   cancer  from  exposure to
   diesel particulate was  made  using  estimates  for the  potency of
   diesel  particulate  and the  earlier estimates of average  urban
   exposure.   Due  to  the  limited  epidemiological  data  available,
   the  estimate of  the  carcinogenic potency  of  diesel particulate
   was  made  using  a  comparative  potency  method developed by
   EPA.[6]   In  this  methodology,  the relative  potency  of  diesel
   particulate to  known  human carcinogens  is  determined from the
   relative potencies  of  the  compounds in non-human  laboratory
   bioassays and  then  applied to known  human  cancer   risks of the
   human  carcinogens.


     As the size  ar.d  cherr.icsi  corposition of diesel particulate
rr.akes it very effective in boch  scattering  and absorcing light,
a  was developed  to  quantify the reduction in visibility caused
by a.Tt.ient diesel particulate levels in  a  large  number of urban
areas.t5]    The  model  used  the  projections  of  ambient  diesel
particulate  levels  described  earlier,  Beers'  law,  a measured
coefficient  of  extinction  for  diesel  particulate,   and  the
assumption that diesel  particulate  levels were  constant  inside
the  city  radius  and  zero outside  the radius to  determine the
visibility reduction.

     The  effects  of   soiling  due  to  diesel  particulate  are
described briefly  in the Draft  RIA.   Little  physical  data were
found describing the rate of particulate  soiling  or the soiling
of   diesel   particulate  relative  to   that   of  other  types.
However,  due  to  its  black  color  and  oily  nature,  diesel
particulate  may  have  a  disproportionate  effect  on  soiling
compared  to  the effect  of  other types  of  particulate matter.
The  only  quantitative  estimates  of   soiling  were  economic in
nature  and  made  in  Chapter 8  of  the Draft  RIA (Cost-Benefit

     A more complete  description of the methodologies described
above can be found in the Draft RIA and the DPS.[5]

II.  Summary  and Analysis  of  Comments  on  NPRM  Environmental
     Impact and Air Quality Projections

     Numerous  comments were  received  from vehicle  and engine
manufacturers,   public  transit  organizations,   environmental
groups  and  private  citizens,  dealing  largely  with  various
specific  inputs  used  to  project  future  emissions  and  air
quality in  the NPRM  analyses.   Several  of  the issues  addressed
are common to both the  NOx and  diesel  particulate analyses, and
will  be  dealt  with  in  the  first  part  of  this  section.  This
discussion of common parameters will be  followed  by two sets of
discussions dealing  with factors specific  only to  the NOx and
particulate projections, respectively.

     A.    Factors Common to Both Analyses

     1.    Baseline VMT Breakdown

     A   critical  parameter   in   estimating   both   NOx   and
particulate emissions  is the breakdown  of  VMT by vehicle class
in  the  area being  examined.   These  VMT breakdowns were under
study  by  EPA just prior  to  the issuance of  the NPRM.  At that
time, it  was discovered that the VMT  breakdowns  used in the NOx
projections,  which  were taken  from the National Emissions Data
System  (NEDS)[1]  for  selected  SMSAs,  were  quite  different  from
the  "Nationwide Urban"  VMT  breakdown  used in  the particulate
analysis,  which was  developed  primarily from  the Energy and


Environmental Analysis,  Inc.   (EEA)  fuel  consumption model.[7]
At the hearing  following  the  NPRM, a  technical  report  entitled
"Motor Vehicle  NOx Inventories"[8]  was  issued  showing  that the
"Nationwide  Urban"  approach   allocated   a  significantly  lower
percentage  of  total  urban VMT to  heavy-duty  diesel  vehicles
(HDDVs) than  did  the  NEDS methodology.   Investigation  into the
NEDS  method  of  county-by-county  allocation  of statewide  VMT
revealed  some likely inaccuracies,  especially  with  respect  to
an overestimation of  HDDV VMT  in urban areas.  The suspected
overestimation by NEDS was confirmed by  estimates gathered from
local transportation  and  planning  authorities,  which on average
indicated a HDDV  fraction of  VMT  very close to that estimated
using the "Nationwide Urban" approach.[8]

     Comments  received on the base-year VMT breakdown  used  in
the  NOx  projections  and   the  above-mentioned  technical  report
indicated  support  for  the  use  of  the  local  transportation
agency data   from  each of  the  cities being  modelled   for  NOx
emissions.   However,  as  the   technical  report  explained,  local
data were  available  for  only  seven of the  eleven cities in the
NOx  analysis.  The  use  of updated  1981-83  average NO2 design
values (discussed  below)  resulted  in  the introduction  of  three
new  cities  into the  NOx  analysis  for which  no  local estimates
have  been obtained  and  the  removal  of  two  cities for  which
estimates  were  available.  Thus,  local   data are now available
for  only  a  minority of the  cities  being modelled.   To further
complicate  matters,  subsequent  analysis  uncovered   inaccuracies
similar to  those found with   the  NEDS approach  in  two of the
seven available local estimates.[9]
     Therefore, both  the  NOx  and diesel  particulate  projections
presented  in  this  final  rulemaking are based on VMT breakdowns
by   vehicle   -class  developed   using   the   "Nationwide  Urban"
approach,  which are very  similar  to  the  average of  the   local
data  which  are available  and contain  no  known errors.   This
method provides the  flexibility  needed   to  accommodate ongoing
changes in  the  cities being analyzed, yet  addresses  the largely
non-urban  nature of  HDDV  travel  (an  improvement   over NEDS).
Because the   "Nationwide Urban"  approach  has  been updated  to be
consistent  with MOBILES  (the  model  used is  called the MOBILES
Fuel  Consumption  Model)*, the  breakdown  of  VMT  by  class  is
     The MOBILES  Fuel Consumption Model  (M3-FCM)  is  a recently
     developed  model,  similar  in  principal  to   EEA's  model,
     which estimates  nationwide and  urban VMT and  fuel usage by
     vehicle  class   and   fuel  type.   EPA's  model  is  based
     primarily  on  MOBILES  fleet  characterization data  (from
     NPTS  and   TIUS)  and  uses  historic   trends  in  vehicle
     registrations   (from  R.L.   Polk)  to  project  future  VMT
     (mileage/vehicle assumed  to  be constant over  time).  Urban
     VMT  fractions  and gas/diesel  sales  splits   used  in  the
     model are  those presented in Tables A-2 through A-5 of the


 slightly   different   than   that   shown   in   "Motor   Vehicle  NOx
 Inventories";  however,  the  basic  rrethodclogy  and  the  urban
 fractions  of  VMT for  each  vehicle  class  are essentially  tne
 same,  while  only the nationwide  VMT  breakdown by vehicle  class
 differs.   In  particular,  the resulting  HDDV fraction of  urban
 VMT   is  essentially   the   same   as   that  with  the  nationwide
 approach   presented   at  the  hearing  and  the  average  of  the
 available  local  data.    (Annual  VMT   by   vehicle   class,   as
 estimated  by the 'MOBILES Fuel Consumption  Model  and used in the
 final  analyses,  is presented  in  Table A-l of  the Appendix.   The
 urban  fractions  of  VMT used  are  shown  in Tables  A-2 and  A-3 for
 heavy-duty diesel  and  gas vehicles,  respectively.[10]    Urban
 fractions  of LDV and  LDT  travel are assumed  to  remain constant
 over  time  at 0.597 and  0.514,  respectively,  based on  1983  FHwA

      Final estimates  of 1982 urban VMT  breakdown by class, used
 in both  the  final  NOx  and diesel  particulate  analyses,  are
 presented  below:         --        -              -   —  ..      ._

             Vehicle Class              %  of  Total  1982  Urban VMT*

 Light-duty Vehicle  (LDV)                        72.8
   -  Gasoline                                          (71.2)
   -  Diesel                                             (1.6)
 Light-duty Truck (LDT)                          20.5
   -  Gasoline                                          (20.1)
   -  Diesel                                             (0.4)
 Heavy-duty Gas Vehicle (HDGV)                    4.4
 Heavy-duty Diesel Vehicle  (HDDV)          	2.3	
             Total                             100.0

 These  percentages,    applied   to  1962  VMT  totals   and  then
 multiplied by  1982  NOx and diesel particulate emission factors,
 were  used to  develop  base-year pollutant  inventories  for  the
 emissions  projections presented later in this chapter.

      2.    VMT Growth Rates

      A modelling parameter that received  a  substantial  amount
 of comment was the set  of  VMT  growth rates that  were applied to
 base-year  VMT  for  each vehicle class  to  project  future  VMT.
 Specific   recommendations  concerning  the appropriate  levels  of
 VMT  growth were  submitted  by General Motors  (with  support from
 other  manufacturers)  and  DOE   (quoting   EEA-based   figures).
 Comments   were   also   received  from   the   American  Trucking
      Because of  the use  of updated  NO2  design  values  (to  be
      addressed  later  in  this chapter),  an update from  1981  to
      1932 base-year VMT was necessary.


Association  (ATA),  stating  that  future VflT  by HDDVs  will be
reduced   due   to   the   replacement   of   some   conventional
truck-trailer   combinations   with   twin   trailers   (i.e.,  cne
tractor  pulling two  trailers).   Although ATA carr.e  to no  final
conclusion on  an appropriate HDDV growth rate, Argonne National
Laboratory was  cited as a reliable  independent  source.   As the
Argonne  VMT model  (TEEMS)  is  being considered for  use  in the
Federal  government's   National   Acid Precipitation   Assessment
Project,   and   recent  output   of  the  model  was   available,
Argonne's  independent  projections  of  future  VMT  growth are
included for purposes of comparison  in this  analysis.[12]

     In  general,  GM's estimates for  each of the vehicle  classes
are  lower  than  the  growth rates  used  in the  NPRM projections
and  lower  than  those  recommended  by   both  DOE  and Argonne.
Table  4-1  summarizes  the VMT  growth  rates  suggested  by the
commenters  (along with Argonne),  compared  to  the rates  used in
.the  NPRM analyses  and those  chosen for the FRM projections.

     The final  (FRM)  growth  rates  shown  in  Table 4-1  are  based
on   urban  VMT  projections   made   using  the   MOBILES   Fuel
Consumption  Model  (M3  FCM),  calculated from  the  VMT  figures
shown  in Table A-l.  This is  the same model used to develop-the
base-year  urban VMT breakdown  by  vehicle  class.    The  growth
rates  are  nationwide averages for urban  areas across  the U.S.;*
city-specific  growth  rates  were  not determined  for  the  same
reasons  given  earlier  in  the base-year  inventory discussion —
absence  of  local  projections   from some  cities  and  need to
accommodate changes  in  the specific  cities being modelled.

     As  Table  4-1 shows,  the  FRM (M3 FCM)  growth rates  for the
LDV  and  LDT  classes   are  in  basic  agreement   with  Argonne's
independent  projections,  estimating  LDT growth at  a  slightly
higher   level   than  LDV  growth.    The  LDT   growth  rate   is
significantly  lower than  that  used in the NPRM analysis,  which
was   based  on  EEA's  Eighth   Quarterly   Report.[13]     GM's
projections  also show  equal rates for LDVs and LDTs.   However,
their  light-duty growth rates are  significantly lower than the
other  estimates, most  likely  due to  GM's  assumption that  both
LDV  and  LDT VMT growths are primarily a function  of growth  in
U.S.  population.    Although  GM  does   state  that   there  were
adjustments  made  to account  for trends in per-capita  vehicle
ownership  and  in miles driven  by  individual vehicles,[14]  their
approach still  appears  to underestimate future  light-duty VMT
growth  in  comparison  with  independent  projections  from  both
Argonne  and  EEA,  based  on  more sophisticated econometric  models.
      In addition  to urban VMT  growth  rates, nationwide  growth
      rates were also calculated  from the  M3  FCM for use  in  the
      NOx  analysis;  both   the   urban  and   nationwide   growth
      estimates are shown in Table A-7 of  the  Appendix.

                           Tacie 4-1
             Annual  Compound  Uroan  VMT Growth Rates
                         (Percent per  Year)
HDV (overall)
+ 1.7
+ 6.4
— _
Ana Ivsis
+4 .0
+ 2.1
+ 6.7
— —
+ 1.2
+ 1.2
+ 3.6
+ 1.1
DOE Araonne FR.M
-- . +1.9 +1.9
— • +2.3 +2.1
+ 0.6
+ 6.9 — +4.2
+2.0 +2.0
     NPRM —     Based   on   EEA's   Eighth   Quarterly    Fuel
                 Consumption  Model   Report   with  assumptions;

     EPA Interim Analysis— Based on EEA 10th Quarterly Report,
                 with  urban  assumptions  from  TIUS  and  FHwA;
                 1981-1995.                        _
     GM —
     DOE --
Based   on
OBERS   with   assumptions;
Based on EEA data and projections; 1980-1995.

Based on ANL-83N  forecast,  TEEMS; NAPAP likely
to be similar; nationwide estimates; 1980-2000.

Based  on   MOBILES   Fuel   Consumption   Model;


     With  respect  to overall  heavy-duty growth,  GM's estimate
is  based  on 1980 Department of.Commerce  (DOC) OBERS projections
for  future   growth   in  employment  within  the  construction,
manufacturing,  and wholesale  trade industries[15]  and GM's is
again  significantly  lower  than  the  figure  estimated  by  both
Argonne  and  the M3  FCM.   However,  use  of  employment   growth
would  again be  expected  to underestimate growth  in VMT, since
employment  grows  more   slowly   than  economic  output   due  to
productivity   improvements   and  heavy-duty   VMT   should   more
closely  follow  the   latter.   For  instance,   if  GM  had   chosen
growth   in   industry   earnings    (also   included    in   DOC's
projections)   instead  of   jobs   as   an  indicator  of   future
heavy-duty  travel,   the  new  figure  would  be   roughly   3.2
percent/year.[15]   Thus,   the   FRM  projections  appear  quite
-reasonable.                       .

     This  overall  growth rate  for  heavy-duty VMT  must  then be
split   between  gasoline-powered  and   diesel-fueled  vehicles
(HDGVs  and  HDDVs,  respectively).    The  MOBILES .Fuel Consumption
Model  determines  this  split  using  diesel  sales  penetration
rates  developed along  with MOBILE3,[10]  the  contents  of which
were  critiqued  by  vehicle  and  engine manufacturers  and other
interested  parties through  a number of workshops.

     3.     Diesel Sales  Projections

     Manufacturers   (primarily  GM)  recommended  significantly
lower  future   light-duty  diesel  sales  fractions  than those
projected   in   the  NPRM,   suggesting   1995  model   year   diesel
penetrations  of  5 percent and  7  percent  for  LDVs  and LDTs,
respectively.   These  estimates  compare to NPRM  figures  of  11.5
percent and 34  percent,  respectively.

     Future  "light-duty  diesel  penetration  is  difficult  to
predict,  as  the demand for  diesels  is  very  dependent  upon
future  oil prices and  the  availability of diesel engines which
satisfy  consumer preferences.  However,  during  the development
of  the  MOBILES  heavy-duty  conversion  factors,  manufacturers
(particularly   GM)   argued   for   substantial    fuel   economy
improvements  well  through  the  1990's,  indicating  a belief  that
fuel   prices  will   indeed   rise   in  the  future   calling   for
continued  improvements  in  fuel  economy.   Therefore,   to  remain
consistent  with this position,  growth  in diesel penetration —
a  fuel-saving  technique — was  also  projected to occur.  'EPA
raised  this  issue  at  that time,  indicating  that  substantial
vehicle-related fuel  economy  improvements  must   logically be
accompanied by  increasing  diesel usage.   EPA accepted most of
these  fuel  economy improvements  predicted by  the  manufacturers,
which  lower heavy-duty  emissions in the future without  direct
emission  control.   Thus,  to argue  for  low diesel  penetrations

now is  quite  inconsistent  with GM's  position just  a  year  ago
and inconsistent with  fuel  economy improvements  assumed  in tne
derivation of  the heavy-duty conversion factors.[10]

     Therefore,  model year  diesel sales  fractions  used  in  the
FRM  analyses   are  similar   to  those  estimated  in  the  NPRM
projections (post-1994  estimates  of  11.5  and  34  percent  for
LDVs  and  LDTs,   respectively), except  that  pre-1995  estimates
have been reduced to reflect slowed growth  (1990  projections of
5 percent  and  15 percent,   respectively).   However,  to identify
the   impact   of  potentially   lower    diesel  penetration   on
particulate emissions,  a sensitivity  analysis will  be performed
wherein the 1990 penetrations  (5/15  percent)  are held constant
through model year  2000 (results to  be discussed  in the final
section  of  this   chapter).*    A  complete  listing  of   the
light-duty model  year  diesel  sales  factions used  in the  FRM
analyses is provided in Table A-4  of  the Appendix.

     While current  light-duty  diesel  penetration is  relatively
low,  particularly in  light  of  GM's recent  decision  to withdraw
from the market, the 11.5 percent 1995  LDV  penetration is still
realistic given that diesel  penetration jumped from 0.3 percent
in  1977  to   6.0  percent   in   1981  with   only  one  domestic
manufacturer   producing  diesels.   Given this fact,   plus  the
potential volatility of  world  oil prices,  it is not difficult
to project a rapid  increase in diesel sales  if fuel prices were
to  increase dramatically.   Furthermore,  in  the  development of
MOBILES   and   elsewhere,    manufacturers    have   consistently
predicted a continued  need   in the  next  decade  to  improve  the
fuel  economy   of   their  engines/vehicles,  and  EPA's  diesel
penetration rates are not inconsistent with these forecasts.

      In view of  the current (1983)  level of  diesel  penetration
into  the  LOT market — approximately 8  percent  —  and the  fact
that  the  diesel  fraction   of  LDT  sales   has  been  steadily
increasing since  1978,  it  is  apparent  that  LDT diesels  are  a
growth market.   Given this,  GM's  estimate of  7 percent for  1995
seems  unrealistically  low,   particularly since GM  supports the
need  for  future fuel  economy  improvements and  does  indeed
predict growth  in diesel penetration of  all other  markets  (LDV
and HDV classes).   Therefore,  15  percent  is a  more  realistic
lower  limit  for  the  sensitivity analysis,  maintaining  a   best
estimate  of  34  percent diesel penetration  into  the  LDT market
by 1995.
     NOx emission factors for gasoline  and  diesel  LDVs and LDTs
     are  quite  similar.   Therefore,  future NOx  emissions  are
     not sensitive to light-duty diesel penetration.


     GM  also  commented   on   diesel   penetration  of  selected
heavy-duty classes,  recommending  1995  figures of 25 percent and
52 percent  for heavy-duty Classes  III-V and  VI,  respectively.
Although the  NPRM  analyses assumed slightly higher penetrations
for these classes,  the use of MOBILES for  the final  rulemaking
projections implicitly assumes  diesel  fractions consistent with
the heavy-duty conversion factors  analysis;[10]  these  figures,
also  used  as  input  to   the  MOBILES  Fuel  Consumption  Model,
essentially are  in  agreement  with GM's  estimates  (30  percent
and  53  percent  for  Classes   III-V and  VI,   respectively).   A
complete listing of  the final  heavy-duty diesel sales fractions
appears in Table A-5 of the Appendix.

     4.     Heavy-Duty  Conversion Factors

     A  fourth issue  — heavy-duty emission  conversion  factors
— has  been addressed extensively in  the MOBILES  workshops and
documented  in  an   August   1984  technical   report.[10]    No
commenter brought any  new  information to  bear  in this area.  As
EPA  has  made  known  in   the  past,[16,17]  MOBILES  conversion
factors for both HDGVs and  HDDVs are significantly  lower than
those used  in the  NPRM analyses (based on MOBILE2.5).  However,
GM's  contention that  even  further  fuel  economy  improvements
should  have   been   incorporated  (resulting   in   even   lower
emissions)  appears  inconsistent with  their projections  of low
diesel  penetration  into   the   light-duty  markets  and  slightly
lower projections  for  the heavy-duty  market.   Therefore,  the
FRM  analyses  will  continue   to  use  the  MOBILES  conversion
factors.   (The final  MOBILES  conversion factors  are presented
in Table A-6  of the Appendix.[10])

     5.     Validity of Rollback Air Quality Models

     The  final  issue  common  to both  the  NOx  and particulate
analyses  is   the   validity   of  the   "rollback"   approach  to
predicting  future  air  quality,  where  any change  in emissions is
assumed to  translate  proportionately  into  a  change  in  ambient
pollutant  concentrations.  In  submitted  comments,  Ford   (with
support from  MVMA) estimated  that  only  one-fifth  to one-third
of- the  change  in  emissions  due  to VMT  growth,  not  the entire
change,   should  be  applied  to  air  quality   projections;  this
estimate is based  on area source dispersion modelling conducted
by Ford.[18]

     Investigation  into  Ford's urban  analysis  uncovered  some
assumptions which  could have  biased  the results of  the study.
One, the traffic density  (VMT/square mile)  at  the center of the
city was  assumed to remain constant.  While  VMT growth  at city
center  is certainly rrore  restricted than  that at the outskirts,
this  assumption allows absolutely no  consideration  for   urban
redevelopment  nor   roadway    construction   or    improvement.


Fur therrrore,  the  assumed  r.odel  of  VMT  density  forced  post
growtn  xn  VMT to be applied to  the  outer edges of the original
urban  area and  to  areas  even beycnd  the original city  radius,
as the  square of the city  radius  was increased in proportion  to
assumed emission growth.

     Two,   the   choice  of  location  for  the  two  air   quality
monitors,  when coupled with  the  above assumptions, also  appears
to minimize the impact  of motor vehicles.   The first monitor,
located at city-center, would be  primarily affected by the area
just  upward of  city-center,  where  VMT  growth  has been  assumed
to  be  essentially  zero.   The  second  receptor,  located  10   km
directly  downwind of  city-center,  would  also -be most affected
by  emissions  in  areas   again   assumed  to  experience  little
growth.   Monitors  not in  line with  the city-center, which were
not included  in  the study, would  be  expected to experience more
VMT growth than  was assumed to be present  in the more congested
areas,  and would therefore  be  more  likely  to  demonstrate  the
impact  of  motor  vehicle emissions.

     An    uncertainty   present    in   Ford's   urban  dispersion
modelling   is  the  selection  of  only  one  stability  class.,
-"slightly   unstable."   As  no   information was   given   on   the
characteristics  of  this and  other  classes,  it  is difficult  to
assess  the impact this choice had on the results.

     EPA   and  others  have  used  rollback  modelling  to  project
future  air  quality since  the  mid-1970's,  and  EPA  has  long
approved   its use  in  State Implementation Plans'for projecting
compliance.   Validations  of  the rollback model  as  applied  to
carbon  monoxide  and  lead  were  included  in  Chapter 3  of  the
Diesel  Particulate Study;[5]  and the   figures  presented  there
show   a   strong  correlation   between  emissions  and   ambient
concentrations  over a  decade.    While   dispersion  modelling   is
probably  more accurate, it is not  feasible in terms of  expense
or  time in  a  study such  as  this to  evaluate  every city  using
dispersion modelling.   Instead,  a simpler  approach,  such as  the
rollback   model,  must  be  used.    Given  the  apparent  bias  and
uncertainties  in the Ford study, it  would be inappropriate  to
discard  or  significantly  adjust   the   rollback  model  here.
However,   possible  improvements  to  the  rollback approach,  such
as  modified  source  discount  factors,  will  be  considered  and
could  be  incorporated into future modelling efforts  if merited.

      6.     Significance of the Air Quality Impact

     Many comments  were received concerning  the  significance of
the  projected increases  in urban concentrations of  particulate
natter   and  N02   due   to  truck   emissions.    MVMA  and   DOE
questioned  what portion  of  the   future  particulate   ambient
 levels' can  be  attributed   to  diesel   trucks.    The  engine


 manufacturers   also   questioned  whether  the  increases  in   NO^
 levels  warrant  the  standards  that  have   been  proposed.    The
 environmental   interests,  and  most of  the  private  citizens  who
 chose   to  comment,   uniformly  criticized   EPA  for,   in  their
 impression,  setting  standards  designed to hold  emissions  at
 current levels and not  attempting  to achieve net  reductions.

     The  standards  that have been  established for both  light-
 and heavy-duty truck NOx and  heavy-duty diesel particulate  have
 been based upon  requirements  of Congress,  which  primarily  focus
 on  technological feasibility and  not  only  on  environmental
"impact  (the  reader   is  referred  to  the Preamble  to   the  final
 rule).   For example,  with respect to  the particulate  standards,
 the Act  calls for  the most  stringent standards  yielding  "the
 greatest  degree  of   emission  reduction achievable through  the
 application of  technology  which  the Administrator  determines
 will be   appropriate  consideration  to  the
 cost...and  to  noise, energy,  and  safety factors  associated  with
 the application of such technology."   Thus,  the  availability of
 technology   is  the   limiting    factor   —  not   satisfactory
 environmental  quality.

     At the same time, the environmental  impacts  described in
 the Draft   RIA,   and below   in" Section  III  of   this  chapter,
 clearly justify  the  need for the standards being promulgated.
 Without  these  NOx  standards,   urban  NOx  levels would   rise
 significantly  over   current  levels  by  the  early  1990's  in
 low-altitude  areas   and  even sooner in  high-altitude  are'as.
 Even with  these  standards,  growth  in emissions  is only  being
 delayed  until the  late  1990's   at  low altitude  and  there is
 almost  no  delay  of  growth at  high  altitude.   Nationwide  NOx
 emissions from all  sources will  also  grow  substantially by the
 mid-1990's,   even  with s substantial  reductions   from   these
 standards.    The  case   for  the  particulate standards is  even
 stronger, given  the  widespread noncompliance  with the  current
 TSP NAAQS   and that expected with  the PMi 0  NAAQS  (discussed
 later).  Thus, the arguments  that the standards  are  either  too
 lenient  or  too  strict  based on environmental  impact  are  not

      B.    Factors Specific to NOx

      1.    Stationary Sources

      Although   no  comments  were   made   pertaining   to   the
 development   of   the  stationary   source   inventories  of   NOx
 emissions,   nor their projected growth,  these  were reviewed in
 light  of   what   was  discovered   concerning  the   NEDS  county
 specific estimates of  mobile  source VMT.    The methodology  used
 by  NEDS  to determine  their  inventories  for  stationary  source


t:0x were  found  to  be  acceptable. [ 19.20]   Therefore,  tne  IIEDS
inventories  (updated  to  1932)  used  in  the  NPRM  air  quality
analysis will continue to be used here.

     Tne  growth  rates  associated  with stationary   source  NOx
used in the  NPRM were determined for EPA by EEA in 1979.  These
were  based  upon certain  population   and  industrial  earnings
growth  factors  as  determined  by DOC/OBERS in  1977.[21]   These
figures have been  compared  to  those  in  the  1980   edition  of
OBERS,[15] and the growth  factors do not appear to have changed
significantly, so the  same rates  are  being used  here.   A more
detailed  review  of this  issue  will be performed in  the  near
future  as the National Acid  Precipitation Assessment  Program
begins  releasing its  projections.   (The final stationary source
growth rates are presented in Table A-7 of the Appendix.)

     2.    NO;   Ambient   Design   Values   and    Inclusion   of

     A  second  issue  specific to the NOx analysis  is the set of
N02  design  values,  or  base-year  ambient  N02  concentrations,
used  in  the  air   quality  projections  for  selected  cities,
Commenters   (Ford,  MVMA)   recommended  the   use   of   average
concentrations over  a  3-year  period to minimize  the effect of
year-to-year  fluctuations  in  monitored  levels.   This  was  in
fact  already  being  done,  as  interim  air  quality  analyses
conducted  after  completion  of  the  NPRM analysis  (early 1984)
were  based  on  NOZ  design  values  averaged over   the  period
1980-82.   These  design values  are being updated  once  more for
this analysis,  as  design values  for  the years between 1981 .and
1983 are  now available.[22]                        —

     With  the  adoption of updated design  values,  the  specific
cities  that  needed  to  be  included in the NO:  analysis (those
with concentrations at  or  above 0.035  ppm  --  66  percent of the
NO2  NAAQS   of   0.053   ppm)  are  different  from  those  cities
modelled  in  past studies.  (Table  A-8  of  the  Appendix lists the
cities  included  in  past  and  current  N02  analyses,  along with
the NOZ  design  values  used  in  the air  quality   projections.)
Also, as  the monitoring period was  updated to 1981-83, the base
v*»ar  ehanaed to 1982  {the  middle  vear}:  therefore.  all  base
year  emissions  inventories  for  mobile and  stationary sources
(both  discussed  in  previous   paragraphs)   used  in  the   FRM
analysis are now calculated for calendar year 1982.

     As Table  A-8  shows,  California cities were  not  included in
the  NPRM NOx  analysis,  primarily  because  California vehicles
are   certified  under  different   (more   stringent)   standards
promulgated  and  enforced  by the California  Air  Resources Board
(CARB) .   However,  CARB corroiented  that many Federally-certified
(non-Ca'lif ornia)  line-haul trucks  cross  over  California state


 lines   and  contribute  to  NOx  and  particulate   emissions   in
 California cities.   Therefore, CARB  feels  that  the  impact  of
 Federal heavy-duty  engine standards  on California air  quality
 should be  evaluated  in  the  FRM.   This is  reasonable.   Thus,
 CARB's projections  of  NOx emissions  for  the  South  Coast  Air
 Quality  Basin  (SCAB),   which  includes  the   three  California
 cities shown  in  the last column of Table A-8,  are  presented in
 the   final  section   of   this   chapter.*   (The   inclusion   of
 California cities in  the  diesel  particulate  analysis  was not an
 issue, as  all urban  areas across  the  nation  were modelled  in
 aggregate; in addition,   air  quality projections were  included
 for  Los Angeles and San Diego in the DPS[5]  and are included in
 the  aggregate results presented in both the  NPRM and the FRM.)

      3.    NOx Emission Factors

      Some  commenters  recommended  that  MOBILE3   NOx  emission
 rates  be  used  instead of those in MOBILE2.   This update  was  of
 course made,   beginning  with  interim  analyses conducted  while
-the  NPRM was  being  reviewed in  early  1984. [23}  For  use here,
 the  MOBILES  inputs  for post-1987 model year LDTs  and HDEs were
 updated to apply specifically to  the  following two  scenarios:
 1) • a  "base case,"  which represents no  further  control  of motor
 vehicle NOx  (2.3 g/mi  and  10.7  g/BHP-hr  for LDTs  and  HDEs,
 respectively), and  2) a  "controlled case," which  evaluates  the
 effect of  the final  standards promulgated  in this  rulemaking
 (1.2  and   1.7  g/mi  for  LOT,  and  LDT2,  respectively,   and  6.0
 followed  by 5.0 g/BHP-hr for HDEs).  The emission  rates  used' in
 the  FRM analysis  are summarized in Tables A-9  and  A-10  for  low
 and   high  altitude   areas,  respectively;  only  those  emission
 rates  and  assumptions  that  are  different  from  MOBILE3  are

      It  should  be   noted  that   the   scenario  designated   as
 baseline   (2.3/10.7)   in  the  FRM • analysis differs  slightly from
 the  baseline  scenario presented  in the NPRM or in  MOBILES.   In
 the  proposal,  future  HDDV NOx  emission  rates were  assumed to
 remain at current   levels  (approximately  7.6  g/BHP-hr)  even
 though the standard  was  set  at  10.7.  In  preparing   the  FRM
 analysis,   this previous  assumption seemed unrealistic  in light
 of  the pressure  that a  particulate  standard  would put  on  NOx
 emissions,  so  the   HDDV  rates  were  instead  adjusted  upward
 assuming   manufacturers  would design for the  10.7  standard once
 they  were sure  it  would  remain at  that  level.   Because -the
      Because EPA's MOBILES  program  does  not have the capability
      to  compute  composite  emission  factors  for  California,
      CARB's  NOx  emissions  and  air  quality  projections  were
      incorporated into the analysis.


 heavy-duty gasoline  (HDGV)  rates  are currently  well below  the
 standard  and  the  particulate  standards do  not  apply  to  these
 vehicles,  no adjustments to the  previous assumptions for  HDGVs
 were r.ade.

      4.     Short-Term NO?  Standard

      The  Natural  Resources  Defense  Council  (NRDC)  commented
 extensively  on   the   need  for  a   short-term   (3-hour)   N02
 standard.    The   NAAQS   for  NOZ  is   currently  under  Agency
 review.   The Agency is currently involved  in  extensive  research
 concerning  the  potential  need for  such  a  standard.   For  the
 time being,  however,  it   is   EPA's  opinion  that  the  current
 annual'  standard  for  N02  provides  adequate  protection  against
 both  long-  and  short-term  health  effects  associated   with
 NOZ .  As   the  basis  for  the  standards  being  promulgated  is
 technological feasibility,  and  not  the  limit of  environmental
 need,  the  existence  of   a short-term NO2  NAAQS  should  not
"affect  this rulemaking, except to  further justify  the  controls
 being implemented.

      5.     Ozone and Acid  Precipitation

      Another issue  specific to the NOx analysis  is the  effect
 of  NOx   reductions   on  urban  ozone   and  downwind   sulfate
 concentrations.  GM (with support from  several other commenters)
 contends  that  a  decrease  in  NOx  emissions will cause  urban
 ozone  and  downwind   sulfate  levels  to  rise.   NRDC,  however,
 disagreed  with  GM's  view  on  ozone formation,  citing  various
 sources  who  maintain  that NOx control  (as well  as HC  control)
 is  essential  in  the  reduction  of  ozone  levels.   NRDC  does
 suggest  that an increase  in urban  NOx emissions may lower ozone
 levels   locally  (as GM  contends),  but  it will  also result  in
 increased  ozone  concentrations   downwind  of  the  higher  NOx
 emissions, merely delaying peak ozone formation.

      The  exact  relationships  between  NOx and  the  other  two
 pollutants  are  rather complex  and  have  been the  subject  of  a
 fair amount of  controversy over  the  past  decade.   Numerous
 factors    play   a   role   in   these  relationships,   including
 (specifically  for  ozone)   the  ratio   of  HC  to  NOZ   ambient
 concentrations,       meteorological       and       topographical
 characteristics   of  the  area,  spatial  location  of   the  NOx
 reductions,  and  others.    Therefore,   the  relationships  could
 differ   from  one  urban area  to  another.   In  addition,  existing
 scientific   studies   of   the   NOx/sulfate   and   NOx/ozone
 relationships are limited,  and their  results  have  not  yet been
 adequately  reviewed  or  accepted by  the  scientific  corr.r.unity.
 An  EPA-sponsored  study   of   the   NOx/ozone  relationship  is
 currently  underway;  however,  the results  are not  yet  available
 and, in any event, are unlikely to support  net  increases  in NOx

      As will be shown  in  the final  section of this chapter, the
 NOx  standards  promulgated  in  the  final   rule  will   prevent
 substantial growth  in NOx  emissions beyond current levels, but


will not significantly  decrease  NOx emissions between  1982  and
the  1995-2000  time  frame.   Therefore,  since a  large  reduction
in total NOx  is  not an  issue  here, no substantial  increase in
ozone or  downwind sulfate  is  suggested.   Also,  the possibility
of reducing ambient  ozone or  sulfate  concentrations by allowing
NOx emissions  to  increase  significantly is not now considered a
viable  long-term  option.   To   allow  concentrations   of  one
dangerous  pollutant  (N02)  to  increase  in  hopes  of  lessening
other pollutant  levels  would  not  appear  to be  wise.   Instead,
EPA  will  most  likely address  the need  for  further ozone  and
sulfate control  in  the context  of HC control   strategies  and
acid precipitation policy.

     Several   other   comments  were  received   concerning  the
relationship    between    truck    NOx   emissions   and    acid
precipitation.   The general  comment  from the  manufacturers is
that controlling truck NOx  emissions  is an inappropriate  way to
control  acid   precipitation,  since it  only  represents a  small
percentage of  emissions  producing  acid precipitation.  GM also
cites  the  fact  that  nitrate  is much   less  acidifying  than

     Environmental  groups  (specifically  NRDC)  were,   in  their
words,  appalled   at  the   lack  of   any   reference   to   acid
precipitation  in  the Draft  RIA.   They  recognize  that,  overall,
SO2  has more  importance  in  terms of  acid  precipitation,  but
insist that NOx cannot  be ignored.  NRDC  refers  specifically, to
the  Western   U.S.,  where  NOx  contributes   over  half  of  the
acidity in  precipitation,   and  to  such seasonal  events  as  the
spring snowmelt, where nitrates dominate the acidity.

     There has been a great deal  of  controversy over  acid rain
in recent years  as  to its  causes  and effects,  primarily  due to
the  complexity  of  the  issue  and  the  lack  of  substantial
clear-cut  data  on   the  subject.   Although   knowledge  of  acid
precipitation  is incomplete,  it is clearly  becoming  a problem
over widespread areas of the country.

     Although  NOx emissions  contribute only  about a  third of
all   acid   deposition   in   the   east, [24]  they   may   have  a
disproportionately  higher  impact   in  terms   of   their  effects.
For  example,   nitric  acid  tends  to become  concentrated  in the
winter  snowpack and  is  then  released  during the  spring  thaw,
creating  episodic  "hot spots"  of acidity  which unfortunately
tend  to coincide  with   the spawning  period  for  fish and  the
beginning of new growth  for plant  life.[24]

      In  contrast to the east, NOx is  the predominant  acid  rain
precursor  in   the western part of  the  United States.   This is
due   primarily  to   the  use  of   low-sulfur  coal  in  western
powerplants,  which   results  in  only 20 percent  of  annual U.S.
SO2   emissions  being  produced  in  the  states  west  of  the
Mississippi River.[24]


     Also,   while  SO.-   is  primarily  emitted  from  stationary
sources,   NOx  production  is  a  joint  mobile  source/stationary
source problem.  As will  be shown  below  in Section  III,  motor
vehicles  are responsible  for  almost  one-third of nationwide NOx
emissions.    In  the absence  of  further  controls  for LDTs  and
HDEs,  nationwide  NOx  emissions  will  increase  by  23  percent
between  1982  and  2000.   With  these  controls,   emissions  will
still  increase  14  percent by 2000,  but will  be  8 percent lower
than   uncontrolled  levels,   which   represents   a  significant

     Thus,   at   this  time,  it  cannot  be  concluded  that  motor
vehicle NOx controls have no effect on  acid precipitation.   Nor
can  it be  stated that such  controls will  play  a large  role in
acid precipitation control  policy.   Identification of  the  most
appropriate  role for  motor  vehicle NOx  control  must wait  for
the  completion  of  the  in-depth  evaluations  of  the  formation,
transport,   and  welfare  effects  of  acid  deposition which  the
Agency has underway.  However, as the  health  effects associated
with  both  current  and  future NOx  emission  levels  justify  the
need for these standards,  this rulemaking  need  not wait  for the
completion of the acid deposition studies.

     6.    Visibility Effects

    • NRDC  commented that  NOx  can  play  a  part  in  visibility
degradation,  either  in   the   form  of  NO2  gas  or   nitrate
aerosols.    They  indicate  that   31   percent   of   the  light
extinction  attributed  to  mobile  sources  in Denver  in  1980  was
due  to motor vehicle NOx emissions.

     The  effects of  NO2   on  visibility  were  examined   in  the
review of the NAAQS for nitrogen  oxides.[25]   Th6 conclusion by
EPA   at   that    time  was   that,   although   N02   does   have   a
visibility  impact,  the improvement  in  visual air  quality  to be
gained by   reducing N02  concentrations was uncertain  at  best.
Due  to this uncertainty,  NOx-related  visibility  impacts  have
not  been  considered  in  this  rulemaking.   However,   as  the
standards  being  promulgated  in this  rulemaking will  reduce
future N02  levels in the  atmosphere from  what  they would have
been,  to  the extent N02  affects visibility,  future visibility
should improve.

     C.    Factors Specific to Diesel Particulate

     1.    Health Effects

     NRDC,  along with  other  environmental groups,  took issue
with how EPA  characterized  the  health effects  due to diesel
particulate matter.   They  agreed with the EPA's  statement that
the   cancer  risk   due    to   diesel   particulate   matter   is


"significant,"  but  emphatically disagreed with EPA's  assessment
of  this  risk as  "small."   NRDC also stated  that "the proposal
notice  makes no  mention of  the  non-carcinogenic health threat
from  fine  particulate  emissions."

      On   the   other   hand,   GM   and  the   American  Trucking
Association   (ATA)  questioned  the  adverse  health  effects  of
diesel  particulate emissions.   Citing  studies  by  the British
Medical  Research  Council  on  London bus  garage  workers,   the
conclusions  of  the  National  Research  Council's  Diesel  Impact
Study committee  and   some  of  their  own studies,  GM  concludes
that   there   is  no   definite  evidence   to   implicate  diesel
emissions  as a "serious  cancer hazard."  ATA  feels that since
"available   evidence   does   not   indicate  that   diesel exhaust
particles  cause  human cancers,"  any reference  to such "should
"be  removed from the  record."   They also question EPA's  use of
relative   potency  analysis  in  determining  the   cancer  risk
associated with diesel particulate  matter.

     "The  non-carcinogenic effects  of diesel  particulate matter
were  detailed  in both the  draft  RIA  and  the  DPS. [5]   These
effects   are compared  to   the  effects  for  other   inhalable
particulate   matter    (PM10,   particulates    less    than  • 10
micrometers  in  diameter),  which,  as  opposed to  TSP,  appear to
be  most directly  related  to adverse non-cancer  health effects.
Based  on   the  available   data,   no   clear   differences   in
non-carcinogenic  health  effects  between ambient PM|0 and f.ine
diesel  particulate matter could be determined,  though there is
some  possibility  that diesel particulate may  be somewhat more
hazardous.   Thus, when  considering overall  health  impact,   the
effect  of diesel  particulate  control  on PMio  levels was used
as  the primary indicator.   As the commenters  submitted  no  new
data  to the  contrary,  this  finding  must stand.

      The  carcinogenic health effects associated  with  the diesel
particulate  matter were also detailed extensively in  the Draft
RIA  and  the  DPS. [5]   The  studies  on  the  London  bus  garage
workers were reviewed in the  DPS and  analyzed independently by
the EPA's Carcinogen  Assessment  Group.   Flaws  in the  design of
these studies  caused them  to  be  disqualified  from further
consideration  in  the DPS,  and no  new  information  has  been
brought   to   light  to   change  that  determination.   Another
epidemiological study is  currently  being  conducted by Harvard
University to evaluate the  possible effect  of diesel  exhaust in
U.S..  railroad  workers.  This study, referred  to  by NRDC,  is
described in the DPS,  and will be  reviewed by EPA  upon   its

      EPA  did  base  its determination  of   the  potential cancer
potency  of   diesel    particulate  upon   a  comparative potency
analysis  that assumes that  the relative  results  of  lower animal


testing   can   be   extrapolated   co   hunans.    While
eoidemiological data are definitely  preferred,  this approach is
not  feasible   until   a   reliable   epidemiological   study   is
available.   Until  then, the  relative potency  analysis remains
the most reliable.

     With  respect  to  the  estimated  cancer  risk,  the approach
was  taken  to   objectively  state  the risk  and  compare  it  to
others experienced by the populace.   Given that  the risk stated
is  a  lifetime risk  for  exposure   to  1995   ambient  levels  of
diesel particulate,  the risk does  not  stand  out and  call  for
control  beyond  that  which  is  technologically  feasible  for
diesels.  However,  at the same time,  the  risk  is not negligible
and does support the need for some degree of  control.

     There  was one additional   comment  on  EPA's  use of  the
proposed   PMio  NAAQS   to   assess   the  effect   of   diesel
particulate  emission   control.    MVMA   feels   that   "it   is
completely   inappropriate   for   EPA  to  anticipate   a   PMio
standard, which has not been  promulgated."   They  cite this as
an act of Mpre-judgment and a compromise of free  ideas."

     The  proposed  standards for  PMio appear  in the  March  20,
1984 Federal Register,  but have  not yet  been promulgated.  Use
of this  proposed  NAAQS was  thought   to have  provided  the most
appropriate  means   of  demonstrating  the   impact  of  diesel
particulate  control on human  health,  as  the  change  to  PM]0
from TSP was proposed  to  more properly  force  control on those
particles   affecting   health.    The  diesel   standards  being
promulgated could just  as  easily have been based, on the current
TSP   standards.    Justification   of   the   light-duty  diesel
particulate  standards  was  based  on  the TSP   standards,  and
noncompliance  with  the TSP  NAAQS   is  projected   tc  be  rrore
widespread  than with  the  PM,0   standards.*    Thus,  use  of  the
proposed  PMio  standards   provides   another   perspective   from
which  to assess the need  for particulate control  and does not
affect  the  result:   diesel  particulate   control  is   justified
environmentally.  The  aspect  affected is  the precision to which
that need, and  the effect of control, is  identified.

     2.    Visibility Effects

     Several   comments  were   received   pertaining    to   the
visibility  impacts  of  diesel particulate matter.  Based upon  a
study  of  four  cities,  GM concluded  that  no  significant impacts
     Between  105  and  329  counties  are  projected  to  be  in
     non-attainment  of  the  proposed  PMio   standard,   compared
     to 300-525  counties  estimated to be in  non-compliance with
     the current TSP standard in the  1987-89  timeframe.[26]


on    visibility   due    to    increased    diesel    particulate
concentrations  will  occur  except  under strict NOx controls (1.0
g/mi  for  LDV,  1.2  g/mi  for  LOT,  and 4.0  g/BHP-hr  for  HDE) .
They  appear  to  have set a 5  percent  reduction  in  visibility as
the  cutoff   for  "significant   impact."    NRDC,   the  Colorado
Department  of  Health,  and  several private  citizens  mentioned
their concern about  visibility,  especially in the  Western U.S.
NRDC  emphasized that  the  reductions  in  visibility  given were
only  averages,  and  that on many  days  the effect could  be much
worse than indicated.

     The  methods  by  which   the  EPA  estimates  the visibility
impact due to diesel particulate  matter are described in detail
in  Chapter  4  of  the  DPS.[5]   These  estimates  are  highly
dependent   upon   the   projections   of   diesel   particulate
emissions.   EPA and GM differ  substantially  in  this respect as
is  indicated by the analysis of other GM comments  earlier  in
this  chapter.   In  the case  of   the  four cities modelled:  New
York  City,  Los  Angeles,  Washington,  DC,   and Denver, GM chose
not  to  project  any VMT  growth except  for  Denver.   Also,  a
fundamental  difference  lies  in the  value used for  the critical
level  of  contrast  against   background   required   to  determine
visibility.  EPA used a value of  5  percent  at airport sites .for
reasons described  in the  DPS.  - If  similar modelling techniques
are  assumed  (i.e.,  Beers' Law),  GM's value  is  closer  to 0.14
percent, which  is well beyond the  level of contrast discernable
by the human eye.  Correcting  for some of these  differences and
considering   the   NOx    standards   being   promulgated,   the
differences  in the   resulting   estimates  of  the  visibility
impacts can be  readily explained.

     The  projected  reductions   in  visibility  due  to  diesel
particulate  are annual  average  reductions,  and  it  is   likely
that  the  effects  will  be  greater on  some  days   and  less  on
others.  However, the  level  of sophistication of the  model and
input  data do  not  allow  shorter term effects  to  be estimated

      3.    Soiling Effects

     A  few comments  were received concerning  the  impacts  of
soiling due  to  diesel  particulate matter.  NRDC, in particular,
cites  estimates of  economic  costs  due to  soiling  ranging from
hundreds  of  millions  to billions of  dollars  annually.  EPA has
reviewed  the scientific and  economic literature  pertaining  to
soil'ing   from  particulate   matter   in   general,   and   diesel
particulate matter  specifically.  The  estimates  of  the benefits
from reduced soiling due to   diesel particulate control shown in
Chapter  8  of  the  Draft  RIA were  in the  same  range   as  the
estimates   quoted   by   NRDC.    Therefore,   there   is  general
concurrence  on  this  issue.


III. E.T.I ss ions/Ai r Quality Projections

     Both in  response  to  the  comments  analyzed  in  Section II
and  co  part  of  the  ongoing  process  of  re-evaluation  and
improvement  of  EPA's  modelling efforts,  EPA  has   revised  its
projections of  future  NOx and  diesel particulate emissions and
air quality.   Several  of  the  input  parameters to  EPA's models
were revised with  the  adoption of MOBILES,[27] and as mentioned
earlier, the comments received  on  the emissions and air quality
model   inputs   were   also   given   full  consideration  in  the
development of final estimates  for each parameter.

     This final  section of  the chapter presents  these revised
projections, based  on  EPA's current  best  estimates for each of
the various input parameters.   Section  A will  deal  with the NOx
projections, followed  by  a discussion of the diesel particulate-
analysis in Section B.   In both analyses,  the methodologies and
inputs  are  the same  as those used  in the  NPRM analyses, except
for the  input  changes  discussed in Section  II  (and detailed in
the Appendix).   For  information on  the methodologies used, the
reader is referred to Section I of this  Chapter and also to the
Draft RIA and the DPS.[5]

     A.    NOx Analysis

     Projections  of   future  NOx   emissions  and  related  air
quality  both  with  and  without  the  promulgated  LOT  and  HDE
standards are  presented below.   First, the  NOx analysis focuses
on  emissions  in key urban  areas  (low-a-ltitude,  high-altitude,
and  California),  and  then  moves  to  projections  of  nationwide
NOx emissions.   The  third part  of  the NOx  analysis  deals  with
the  impact  of  future  emissions  on  ambient  NO,  levels  in the
urban  areas  of  concern,  and   a  final  section   offers  EPA's
conclusions on the need for  future NOx  controls.

     1.    Emissions in Key  Urban Areas

     As  mentioned above,  the  first  part  of  the  NOx  analysis
focuses  on  the  ten  urban  areas  shown  in  Table   A-8 of  the
Appendix,    consisting   of     eight  ,  low-altitude   and   two
high-altitude  cities.   Also, CARB's  projections  for  the  three
California  cities  shown in  the table (all  located in the South
Coast Air Basin) are included in this discussion.

     Table  4-2  presents  base-year,  and  future  "NOx   emissions
inventories  for  the   low-  and  high-altitude  cities  under two
future  NOx  standards   scenarios.    "Base  case"  represents no
further  control  of NOx,  with   a LDT  standard  of  2.3 g/rni and  a
HDE standard of  10.7 g/BHP-hr.   The  "controlled case"  refers to
the  NOx standards  being  promulgated — 1.2 g/mi  and 1.7  g/.-ni
for   the  LDTj   and  LDT2   classes,   respectively,   and  HDE


                           Table 4-2

           Base-year and Future'Urban NOx Emissions*
(1000 tons/year)
Eight Non-California Low-Altitude Urban Areas
— 291
Two Non-Cali

f ornia


* *
Urban Areas**

** 20.3
     NOx emissions  do not  include  those from  stationary point
     sources, due  to  limited  air  quality  impact  relative  to
     ground-level sources.
**   Includes  the   eight   low-altutude  and  two  high  altitude
     SMSAs  listed in Table A-9  (FRM column).
***  Numbers in parentheses represent reductions from base case.


standards of  6.0  g/3H?-nr  in 1958, followed  by  5.0  g/E-iF-hr in
1S91.  Stationary area and off-highway  source NOx emissions are
included in the category  "Others."   In these urban projections,
staticr.ary point  source  emissions are  not  included  because of
their  relatively  low  air quality  impact  per  ton  compared to
that of ground-level sources.

     As  shown,   total  baseline  NOx  emissions  in  the  eight
low-altitude urban  areas  are expected  to grow  by seven percent
between 1982 and  1995,  with an  overall increase of  16 percent
by  the year  2000.    As  in  the  NPRM,  the  largest  increase is
projected for  the  HDDV  class,   with  year  2000   emissions  more
than  double  the   1982   levels.   LDT  emissions  increase  by
approximately 15 percent,  while  HDGV  and  LDV emissions decrease
without further control.   (Shown graphically  in Figure 4-1.)

     The  effect of  the  final   standards  for LDT  and  HDE NOx
emissions in  these  eight  low-altitude areas  is also  evident
from  the  projections in  Table  4-2.   As shown,  controlled NOx
emissions are  estimated  to  be  approximately 10  percent   lower
than  the  base case  in  1995, and  13  percent lower  in  the year
2000.  These reductions  due to stricter LDT  and  HDE  NOx control
result  in total NOx  emissions  (including  those  from stationary
area  and  off-highway sources)  staying fairly constant through
the  year  2000.   Total emissions decrease by 4  percent in  1995
relative  to  1982,  and are  roughly 2  percent higher  than  base
year  in 2000,  with motor vehicle  emissions  18  percent lower in
1995  and  15  percent lower in 2000  (with  respect  to   1982
emissions).   (See Figure 4-2.)

     As  shown  in  the  bottom  portion  of  Table   4-2,  future
emissions growth  in  the  high-altitude areas  is  projected  to be
much  greater  than  in the low-altitude cities.   The difference
is not VMT growth, as the same  national average  rates were  used
for  both  low  and  high-altitude   areas;  instead, growth  is higher
because the 1.0  g/mi NOx standard on  1981  and  later model  yea-r
cars  (LDVs)  and  the 2.3  g/mi   standard  on  LDTs  (beginning in
1979)  did  not  have  as  great  an   impact  on  high-altitude
emissions as  they did at  low altitude.  This is  due  to the  fact
that   pre-control   emission   rates   , for   LDVs   and   LDTs  in
high-altitude areas  were lower  than  those  in low altitudes but
controlled  levels  are about  the same.  Therefore,  the smaller
impact   of   existing   light-duty  controls   on  high-altitude
vehicles does not outweigh the future  VMT growth, as it does in
low-altitude  areas.    This  is  shown  in  Table  4-2,   where
base-case LDV emissions  show a  decrease between  1982  and  1995
in the  low- altitude  areas,  but  stay  the same in  high altitudes.

     Overall,   total  baseline   NOx   emissions  in   the  two
high-altitude areas  are projected  to  grow  by 25  percent between
1982  and  1995  (shown in  Figure 4-3),  compared  to 7 percent in

                   Figure 4-1



 S  700-
J  eoo-l

O  500-
 £  400-

•^  300-


 X  200-



            NOx Emissions Inventory for Eight Urban Areas

                         ;     Low Altitude
                              . Base Scenario

                                                MOBILE SOURCES

                                                          OTHER SOURCES

                                                       CD MODE

                                                       Z2 LOV

                                i. -.'Ore: •<-

S  700-
|  600-)

O  500-
 C  40°"

.£  300-

 X  200-



           NOx Emissions Inventory for Eight Urban Areas
                             Low Altitude
                          1.2/1.^6.0/5.0 Scenario
                                                               MOBILE SOURCES
                                                                     G23 OTHER SOu
                                                                     a HDDE
                                                                     E3 LOT
                                                                     EZ1 LDV


*-   90-
o      '



 £   50-

*S   40"

l§   30H
§  .20-


                                 Figure 4-3  .   •       |!
                 Emissions Inventory for Two Urban Areas
                              High Altitude          f
                              Base Scenario
                                                               MOBILE SOURCES
                                                                       OS OTHER SOURCES
                                                                       £2 LDV


lev:  altitudes.   However,   the  promulgated  LDT  and  HDE  NOx
standards  have  basically   the  sarr.e  effect  on  1995 ar.d  2COO
emissions in both altitudes.  This is expected  because,  by that
time,  . as  mentioned  above,  the  baseline   (2.3/10.7  standards)
emission  rates  in low  and high  altitudes  are quite  similar.
But  even   with   the  stricter   control   on   LDTs   and  HDEs,
high-altitude  emissions  are  expected  to   grow by   14  percent
between  1982  and 1995,  with  a  23 percent  increase  by the year
2000 (see Figure  4-4).  These figures are  quite  large compared
to the  relatively small changes from base-year levels projected
to occur in low-altitude areas with the added control.

     Projections  of  future NOx emissions  for- the  South  Coast
Air  Basin   (the  Los  Angeles   area)  were   provided  by  the
California  Air  Resources  Board  (GARB)   and  are  presented  in
Table  4-3.   GARB  examined  three NOx  standards  scenarios  for
Federal  line-haul  (Class   VIIIB)  diesel  trucks:   10.7  g/BHP-hr
(no further control), 6.0  g/BHP-hr in 1988, and finally the 6.0
standard  followed by  4.0 g/BHP-hr  in  1991.*   Although  the
Federal  standard of  5.0 being promulgated  in the final rule was
not specifically examined   by  CARB, sufficient  data was provided
to  interpolate  between  scenarios.   All  scenarios  assume that
only Federal line-haul trucks will cross  into  California  (i.e.;
none of  the lighter classes).

     As  Table   4-3   shows,  total   NOx   emissions  (including
stationary point sources)   in the  SCAB are  projected  to be lower
than  current  levels  in the  year 2000,  regardless  of  Federal
control.  However, based on the California State Implementation
Plan  (SIP),  total   SCAB   emissions  must   be  at  or  below  895
tons/day  in  order   for  the  cities  in  the   basin  to  be  in
attainment  of  the NOZ  NAAQS.   CARB  projects  attainment  to be
achieved   sometime   in    the    late   1330s,    but   projects
non-attainment by  2000  due to  growth unless Federal (and thus
California)  engines  are certified at  4.0 g/BHP-hr.   However,
(though  not  modelled  by  CARB)  a  Federal   standard  of  5.0
g/BHP-hr may  result  in only  marginal non-attainment,  based on
evaluation  of  the relative emission totals presented  in Table

     2.    Nationwide Emissions

     In  addition  to   evaluating  the  effect  of   the  final
standards  on  NOx  emissions  in  these  specific  low-altitude,
high-altitude  and  California  urban  areas,  the impact  on total
     Due  to  provisions  of  California's   waiver   from  Federal
     standards,  this  Federal  truck scenario  also  assumes   the
     reduction   of   California's  standard   from   5.1  to   4.0


                           Tacle  4-3

                  South Coast Air Basin (SCAB)
                   NOx Projections (tons/day)




2000 Federal
HDE Std. Scenarios
Stat. Area


Degree Above
NAAQS Attainment
Level (%)***  13          6.2          1           -1
*    The 6.0/5.0 Federal  scenario was not examined by CARB, but
    - was estimated by EPA based  on  CARB's data; represents very
    • rrisrgir.3l nor.attainrrent of NAAQS-
**   Totals are up  to  2 percent greater  than  those  provided  by
     CARS,   due   to  round-off   error   in  recombining   source
***  The California SIP estimate is that NOx  emission levels  at
     or below approximately  895  tons  per day  are  necessary for
     SCAB  attainment  of  the  NO2  NAAQS.   Based  on  this,  the
     6.0/4.0 Federal standard,  which  would be accompanied by  a
     reduction  of  the   California  standard   from  5.1  to 4.0,
     allows the  SCAB  to  stay  in attainment  in 2000.   (Initial
     attainment is  projected  for the late 1980's, regardless  of
     Federal control.)

Source:    California   Air   Resources   Board,  Mike   Sheible,
           January 22,  1985, phone conversation.


nationwide  NOx  emissions   was  also  determined.   This  larger
scale  analysis  can  be  especially  useful  in  evaluating  the
secondary effects of  NOx  contro-1,  such as acid  rain  formation.
Because nationwide  projections  were not  included in the NPRM, a
brief explanation of the methodology used is in order.

     Projections  for  the  nation  (48 continental  states)  are
made  using   base-year  inventories  from  the National  Emissions
Data System  (NEDS).*[1]  Motor  vehicle  inventories  are adjusted
for  future   VMT  growth  and emission  control using  nationwide
average  VMT growth rates  from  the  MOBILES  Fuel  Consumption
Model  (shown in Table A-7)  and MOBILE3 emission  factor  ratios
for  the various  standard  scenarios.   Current  emissions  from
other  sources  are  adjusted using assumptions  also  shown  in
Table  A-7.[21,28]   In  this  nationwide  analysis,  stationary
point  sources   are  included due  to  the  larger  scale regional
concerns usually associated with secondary NOx effects.   	

     These  nationwide NOx  projections  are  shown in  Table 4-4
and  in  Figures  4-5  and 4-6; these  "base" and "controlled" cases
refer  to  the  same  standards  scenarios  described  earlier.   As
shown,  without  further  LOT and  HDE  control, total  nationwide
NOx  emissions are  projected to grow by 13 percent  between 1982
and  1995,  with  a  23  percent  increase  by  the  year  2000.
However, with the  final  LOT and HDE  standards in  place,  growth
during  the   same periods  is estimated  to  be 6  and  14 percent,
respectively,  or  an  overall   reduction  of  6-8  percent  from
future uncontrolled emissions.

     3.    Air Quality

     Using  the  rollback model  and  input  data described  in the
Draft  RIA  and Section  II   above,  the effect  of the  final NOx
standards on ambient  N02  concentrations  was  evaluated for the
eight  low-altitude  and two high-altitude urban  areas mentioned
earlier.  Table 4-5  presents   the  results  of  this  evaluation,
comparing  projected  N02   NAAQS  attainment  status  under  both
the  promulgated standards  and  the baseline case.   Because the
rollback  approach  was used, the  percent  change  in  ambient NOZ
concentration  tracks  the  change   in  NOx emissions   (excluding
point sources), which  have  already  been described above.
     Because   the  NEDS  weaknesses   exist   primarily  in ' the
     apportionment  of VMT  to  individual  counties  and do  not
     apply  to  statewide  totals,  the  methodologies  used  to
     calculate  nationwide NOx  inventories  are  appropriate  for
     use in this  part  of  the  analysis.


                      Table 4-4

            Total  Nationwide NOx Emissions
            	(1000 tens/year)	

-- -" HDDV
~-~-- (subtotal)
Stationary Area
.->.; Total . .
2,204( 0%)*
241( 0%)
3,342( 0%)
2,677( 0%)
12, 583 ( 0%)
24,356( 6%)
2,422( 0%)*
241( 0%)
3,478( 0%)'
3,029( 0%)
13,776( 0%)
26,217( 8%)
Figures in parentheses indicate reductions from base case

                              Figure 4-5
                       Base Scenario: 2.3/10.7
O  15000
.82  10000


Z   5000H
                                                                     STAT PT
                                                                  KH OFF-HWY
                                                                  C3 COMBUST
                                                                  £3 STAT AREA
                                                                  CD HDDE
                                                                  ED HDGE
                                                                  a LOT

                                   Figure 4-
                    Controlled Scenario: 1.2/1.7; 6.0/5.0
    15000 H
 £  10000-
     5000 H

                                                  STAT PT
                                               EH OFF-HWY
                                               CS COMBUST
                                               BS3 STAT AREA
                                               CD HDDE
                                               E3 HDGE
                                               C3 LOT
                                               CZ3 LDV


                            Table  4-5
           Average Percent  Change  in  NOx  Emissions  and
      Ambient  NO>  Concentrations  from the Base Year  (1982)*

                              Eioht Low-Altitude Areas**
Base Case:

Controlled Case:
(1.2/1.7; 6.0/5.0)
Base Case:

Controlled Case:
(1.2/1.7; 6.0/5.0)
1990             1995             2QOO

 -1               +6               +16

 -6               -5       '         + 1

	Two High-Altitude Areas**

+ 13

 + 9

+ 26

+ 15

 + 39

 + 24
*    Stationary  point  sources  are not included  in  the emission
**   Negative  value  denotes  a decrease;  positive  value denojtes
     an increase.


     Table  4-6  estirates  the number  cf Standard  Metropolitan
Statistical Areas  (SMSAs),  or urban  areas,  projected  to  oe
above  the  ambient  NO,   standard   of  0.053  ppm  in  several
pro^ec'ion  years.   It  should  be  noted  that  actual  number  of
non-attainment areas shown is not  to  be taken  as  absolute,  as
projections of this  type  are  difficult  to  make.  Rather,  the
relative number of  exceedances is  more  appropriate as  a  means
of  evaluating the   relative   impact  of   a   particular  control
scenario.   As shown, two of  the  three non-California  areas fall
into attainment  with the  final   standards  in  place,  with  the
three  California   cities  predicted  to   be  in  only  marginal
non-attainment in the year  2000.

     4.    Conclusions      --

     It is against the  background of the above projections that
EPA must  evaluate  the  comments by  manufacturers that  there  is
insufficient  need  for   NOx  control  to  justify  the  proposed
standards   for  light-duty  trucks  and  heavy-duty  engines.   Even
with   the   revised   input   data  that  project   lower   future
emissions,  overall  growth  in  future  NOx is  still  projected  to
be significant for both the nation as a whole  and for  the urban
areas  of  concern.   The  same basic  need  for  further NOx control
demonstrated  in  the  proposal  still  exists,  and  current action
is  necessary  if   future   problems   are   to  be  dealt   with

     The  statutory  provisions of  Section 202(a)(3)(E)  allowing
EPA  to   relax   the  NOx  standards   based   upon  air  quality
considerations  place  a   positive  burden   on  the  Agency  to
substantiate  a lack of  need  for  more stringent  levels.  Based
upon its projections of  future emissions and their telationship
to  both  the  attainment of   the  National   Ambient Air  Quality
Standard  and  to   other  actual or potential  secondary  impacts,
EPA  finds  it  impossible  to make  such  a  statement at  this time.
Therefore,  the  standards  promulgated  in  the  final   rule  have
been    developed    under      the    provisions    of    Section
202(a)(3)(B)-(D),   which provide  for  setting  standards  based
upon those  levels which do  not   increase cost  or decrease fuel
economy to  an excessive and unreasonable degree.

     B.    Diesel Particulate Analysis

     Revised  projections  of   diesel  particulate  emissions  and
related impacts are  presented in  the following paragraphs.  The
analysis begins with urban emissions projections  both  with and
without the promulgated HDE  control,  followed  by  a  discussion
of  the impact of  these  diesel particulate  emissions  on  urban
air  quality.   The   final  section  deals  with  the health  and
welfare impacts of  diesel  particulate exposure,  including both
non-cancer   and   carcinogenic   health   effects,   visibility


                            Table  4-6

           Number of  SMSAs  Projected  to  Exceed  the  N02
          	Ambient Air  Quality  Standard	

                     1984     1990     1995     2000

Base Case: (2.3/10.7)

Low Altitude           0011
High Altitude          0122
California  •           I*       0_        Q_        3_
  Total                1136

Controlled Case:  (1.2/1.7;  6.0/5.0)

Low Altitude           0000
High Altitude          0111
California     ~  ' '     1*       0    -    0        3_
  Total                1        1        1        4
     Los Angeles  is the  only  SMSA currently  in non-attainment
     of the NOj NAAQS.


reduction,   and   soiling.    Unless   specified,   the   analyses
presented below  utilize the rr.ethodology  outlined in  the  Draft
RIA,  as modified in Section II  above.

     1.    Urban Emissions

     Unlike NOx,  diesel particulate is modelled for  urban  areas
across  the  nation  in  aggregate,  without  focus  on  particular
cities.  This   is  done  because  violation  of  the  NAAQS  for
particulate is more  widespread than  it  is for NOx.   The  final
diesel particulate emissions projections  are  presented in  Table
4-7.    The  two  future  scenarios shown differ only in  the HDDV
standards assumed; for light-duty diesels,  the  standards that
are currently set  to  come  into effect with the 1987  model year
— 0.20 and 0.26  g/mi  for  LDDVs and LDDTs, respectively —  are
assumed.  The "base"  scenario   represents  no  further  control of
HDDV particulate  emissions,  assuming uncontrolled  emissions at
0.70    grams   per    brake-horsepower-hour   (g/BHP-hr).     The
"controlled"  case  is  based   on   the  HDDV   standards   being
promulgated in  this  rulemaking —  0.60  in  1988, followed by
0.25 in 1991.and 0.10  g/BHP-hr in 1994.   (Urban  diesel  buses
will be subject  to the 0.10 g/BHP-hr standard in 1991.)

     As Table 4-7  indicates, urban diesel particulate emissions
are projected to  grow  to  twice the  current  levels by  the year
2000 if no  further HDDV controls  are imposed (shown graphically
in Figure  4-7).   It   is  this  HDDV category  that  makes up  the
majority of the  total emissions,  representing  84  percent in
1984 and 63 percent  of the  total in  2000.   (This decrease in
heavy-duty   share  occurs   as  the  diesel  penetration  of  the
light-duty   market  increases.)   Table   4-7   also  includes  a
breakdown  by  class  of the  HDDV  emissions,   which  shews  that
line-haul  (Class  VIIIB)  diesels make  up  almost  half  of  total
HDDV emissions  in 2000.

     The effect  of HDDV and urban bus control is significant,
with  the  combined   1988/91/94  standards  bringing   about  an
estimated 46 percent  decrease  from the base  (uncontrolled)  case
in the  year 2000.  This  level of control essentially prevents
significant  growth  beyond  current  levels,  with  about   an 11
percent  increase  projected  between  1984  and  2000  (see   Figure

     The more   stringent   control  (0.10  g/BHP-hr  standard) of
urban   buses,   beginning   with  199.1    models,   and   of   other
heavy-duty  classes in 1994 is a  substantial  portion of  this
overall  impact  on  emissions  by  the  year   2000.   The  -0.10
g/BHP-hr standard  accounts for 23  percent of  the reduction in
emissions  from  uncontrolled levels.   From another perspective,
if  the 1994 0.10 standard  were  not  implemented and  the  0.25
standard simply  continued  on  through  2000  for both  buses  and


                                   Table 4-7

                 Base-year and Future Urban Diesel ParticuLate
                 	Bnissions (cons/year)*	___^_
                          1995 HDEV Scenarios
                                               2000 HDCV Scenarios
LDDV   . 5,699  (11%)**  13,392  (15%)    13,392  (23%)   19,700  (18%)   19,700  (33%)

LDDT     2,492   (5%)    13,072  (15%)    13,072  (23%)   20,713  (19%)   20,713  (35%)

HDDV   - 45,013  (84%)    61,485  (70%)    30,767  (54%)   68,528  (63%)   18,903  (32%)
Total   53,209(100%)   87,949(100%)    57,231(100%)  108,941(100%)   59,316(100%)
                      Breakdown of HDIV Bnissions  (tons/year)*
1995 HDCV
2000 HDDV
7,541(40%) '
 *     "Best estimate" diesel sales fractions,  shown in Table A-5,  are assumed.
 **    Figures in parentheses indicate percent  of total.

                            Figure 4-
             Urban Diesel Particulate Emissions
                       Base Scenario


J  80H

8  7
            Figure 4-8
Urban Diesel Particulate Emissions
       Controlled Scenario
        HDDE:  .6/-25/-1
          BUSES: A/A
                                                  CD BUSES
                                                  C3 HDDE
                                                  C3 LDDT
                                                  IZ2J LDDV


trucks,   total  diesel  particulars  erissions   in  the  year  2000
would be approximately 33 percent higher  than  in  1932; nowever,
with  the final  more  stringent   standards,  growth  during  this
period, is constrained to an estimated 11 percent.

     The emissions projections presented  in Table  4-7 are based
upon  EPA's  best  estimates  for   the  various  input  parameters;
however, because of  the  difficulty  in  projecting  future diesel
penetration into the  light-duty  markets,  a sensitivity analysis
was  performed.    Instead  of  assuming  that  light-duty  diesel
production will  continue to grow through  1995  (as  in the "Best
Estimate"  analysis),   another  case  was  examined  wherein  1990
levels  of  5 percent  and  15  percent diesel penetration  of the
LDV  and LOT  markets,  respectively,  was  assumed   to  continue
through  the  year  2000.   (Best  estimates of  heavy-duty diesel
penetration — less  difficult to predict  —  were used  in both
cases.)   Results  of  the  sensitivity analysis  are  presented in
Table 4-8.

     As  indicated,   the  use   of   the   lower   future   diesel
penetrations results  in a  29-30 percent  decrease in  light-duty
emissions  in  1995  and a  47-50   percent   decrease  in the  year
2000, in comparison to  best  estimate projections for the same
two years.  With  respect to total  diesel  particulate emissions
under  the  "Low  Penetration"   scenario,  assuming  no   further
control,   growth   between   1984   and   2000   would   still  be
significant  at  68  percent  (compared  to 105  percent assuming
"Best   Estimate   Penetration").    With    imposition   of   the
1988/91/94 standards  on  HDDVs,  assuming low diesel  penetration,
year  2000  emissions  would  be   approximately  25   percent  lower
than  current  levels  (compared to the  11  percent  increase over
current  levels  projected  using  best  estimates  of  light-duty
diesel penetration).

     2.     Air Quality

     The  impact  of  growth in diesel particulate  emissions on
urban  air  quality  is  significant,  as   shown  in  Table  4-9.
Current  ambient  diesel  particulate  concentrations  in  large
cities  are  projected  to grow from an  average of 1-3  ug/m3  to
levels  of  3-7 ug/mj  by the  year 2000  with  no further  control
on HDDVs (using best  estimate assumptions).   With the standards
promulgated    in    the    final     rule,    diesel    particulate
concentrations  in  large cities  will  be  reduced  to  1.5-4  ug/mj
(best  estimates),  a  reduction  to  almost  half   of   baseline

     3.    Health and Welfare Effects

     As  discussed  in the Draft RIA and the  DPS, [5]  exposure to
diesel '  particulate   emissions   has   an   impact  on   these   four


                       Table 4-8

 Sensitivity Analysis of Light-Duty Diesel Penetration
Urban Diesel Particulate Emissions (1000 tons/year;
TOTAL ' '"
53.2 * -
~ 88.0
- 80.2
11. Q
 Standards    scenario:   no   further   HDDV   control    (0.7
"g/BHP-hr).   LDDV  and  LDDT emissions  do 'not  change  with
 heavy-duty  control  scenario.
 HDDV  class  includes buses.


                                      •Sable  4-9

             Effect of Diesel Particulate Control on Urban Mr Quality*
Total Diesel Particulate Concentration (ug/ra^)
1995** 2000**
Ambient Urban Concentrations***
City Population: -
- Greater than 1,000,000
- 500,000 - 1,000,000
250,000 - 500,000
100,000 - 250,000
annual Average Exposure to U.S.
Microscale Concentrations
Roadway Tunnel:
Street Canyon:
Severe ,
On Expressway:
. 23

. .1.0-2.4
15 '
~ - 5.5
0.8-2.2 '
  Beside Expressway
*    Based on best-estimate projections.
**   Control effectiveness is approximately 35% in 1995 and 46% in 2000.
***  Ranges are average values plus and minus one standard deviation.


areas:   1)  non-cancer  health effects,  2)   carcinogenic  health
effects, 3) visibility, and 4) soiling.

     a .     Non-Cancer Health  Effects

     Particulate  matter in  general has  long  been  regarded  as
hazardous  to  human health.   EPA  recognized  this  danger  and
established  an NAAQS  for  total suspended  particulate  (TSP)  as
early as 1971.  As  discussed  in Section  II,  EPA  has  proposed an
ambient  standard  that will  focus  on inhalable particles (i.e.,
those with  diameters  of  10  microns or  less  (PMi0))»  because
it  is this  fraction that appears  to  be  responsible  for most- of
the human health effects associated with TSP.

     As mentioned earlier,  diesel-particulates fall  easily into
the  PMto  category,  as  the  majority  are  classified   as  fine
particulate  (less  than  2.5  microns  in  diameter).  Although  a
large  body  of data  has been  developed  regarding  the  health
effects  of   inhalable   particulate  matter,  research  limited
specifically   to  diesel  particulate  is  relatively  new  and
somewhat  inconclusive.   An  analysis  of   the  available  data
indicates   that,   until  more   is  known,   diesel  particulate
generally should be regarded  as  being equivalent to  other forms
of  inhalable  particulate  matter  in terms   of  the hazard  it
presents to  human  health,  although  there  is  a  possibility it
may be  somewhat more  hazardous. [5]   It  should  be  pointed out,
however,  that  even if  regarded  as  posing  the  same  hazard,
diesel  particulate  is  emitted directly into the breathing zone,
rather  than  from  tall  stacks  that  would  promote  dispersion.
Thus, the potential for human exposure is maximized.

     Two basic concerns exist  with respect to  the  health risk
posed by inhalable particulate in  general.  First,  inhalable
particulates  are  small enough  so  that they are not as readily
prevented by  the  natural  body defenses from reaching  the lower
respiratory    tract,    as   would    coarser    particles.    Fine
particulate  matter  can  penetrate  to the  alveoli,  or  deepest
recesses of  the lungs,  where the  oxygen/carbon  dioxide exchange
takes place with  the  circulatory system.[29]  The body requires
months  or  years  to  clear  foreign matter  from the  alveolar
region,  as   opposed  to  hours  or   days   to  clear  the -upper
respiratory   system.    The  second   concern  is  that  inhalable
particulate  may  be composed of  toxic  materials  or   may  have
hazardous materials adsorbed  onto  its surface.

     The most obvious  non-cancer  health effect  of an inhalable
particulate,  such as that produced  by diesels, is  injury to the
surfaces  of   the   respiratory  system,  which  could  result  in
reduced  lung  function,    bronchitis  or   chronic  respiratory
symptoms.   The hazardous chemicals  that  may be  associated with
particulate  matter  (e.g.,  organic  compounds,  lead,  antimony,


etc.) can  either  react  witn lung  tissue or be  transported to
other parts of the body by  the  circulatory system.  Particuiate
matter may  also  weaken the resistance of  the  body to infection
and   vnere   are   indications   that   it   reacts   adversely  in
conjunction  with other  atmospheric  pollutants.   For  example,
studies in London, New York,  Buffalo, and Nashville  have  found
an  increase  in  the  mortality  rate,  especially among   older
persons,  when high particulate  levels were  accompanied  by high
sulfur dioxide levels.[30]

     From  the  above  discussion,   it  is  clear  that  inhalable
particulate  matter  (PM10)   has  been  linked  directly  with  a
myriad of  adverse  non-cancer  health effects, a-nd  it is based on
this  information  that  EPA  has  proposed  the  NAAQS   for   PMi0.
Also,  diesel  particles   are  all  inhalable  particulate  and,
therefore,  can  potentially  represent the  same  concern.  "This
relationship  can  be  used  to  assess  the  overall benefits  of
controlling HDDV diesel particulate.

     As  stated  in the Draft RIA for  the PM,0  NAAQS,  105-329
counties  are  projected to be in  non-attainment  of the range of
primary  PMio standards  being  considered for  1989.[26]   Even
after  reasonable  non-mobile  source  emission   controls  are
implemented,  numerous  violations   of  the  NAAQS   are   still
projected  to  occur.   As  shown  in Table 4-9,  if no further HDDV
controls  were implemented,  annual  average  exposure   to  diesel
particulate   for   urban   dwellers  would  be   at  a   level  of
approximately 5.5  ug/m3  in the year  2000, or  about  10 percent
of   the   suggested  PMio   NAAQS.    Promulgation   of   the  HDDV
standards  is projected  to  reduce  this  exposure  to   about 3.0
ug/m3, therefore playing  an important  role in  reducing  urban
PMjo   exposure.    Furthermore,    ths  resulting   reduction  in
diesel particulate emissions within urban  areas that continue
to  violate  the  suggested  PM,0  NAAQS  will directly  reduce the
non-cancer    health    effects    associated    with    inhalable
particulates  in general.

     b.    Carcinogenic Health  Effects

     A number of studies have concluded  that exposure to diesel
particulate  probably  poses  an additional  risk of  acquiring lung
cancer.     EPA    surveyed    these    studies    and    developed
scenario-specific  risk   factors   for  lung   cancer   incidence,
taking   into  account   the  relative   reduction   of   compounds
producing  the cancer-risk  with respect to  reductions in  total
diesel particulate.[5]   Table  4-10  shows the  resultant cancer
risk  estimates  associated with diesel  particulate for both- the
base-case   and    the  controlled-case   scenarios  along  with
estimated   risks   from  other  known  carcinogens,   shown for
purposes of comparison.


                         Table 4-10

          Comparison of Risks from Various Sources
  Sources of Risk
Estimated Annual Risk
Commonplace Risks of Death
Motor Vehicle Accident
Tornados, Floods, Light-
ning, Hurricanes, etc.
Risks of Cancer Incidence
Diesel Particulate (1995):
Base Scenario 1.2
Controlled Scenario 0.8
Diesel Particulate (2000):
Base Scenario 1.5
Controlled Scenario 0.8
Natural Background Radi-
ation (sea level)
Average Diagnostic Medical
X-Rays in the U.S.
Frequent Airline Passenger
(4 hours per week
Four Tablespoons Peanut
Butter Per Day (due to
presence of aflatoxin)
Ethylene Dibromide
One 12-Ounce Diet
Drink Per Day
Miami or New Orleans
Drinking Water (due
to presence of chloroform)
Lung Cancers:
For Smokers Due to
For General Population


x 10-
x 10'

x 10-
x 10"





























6.2 x 10-'
4.1 x 10'*

7.7 x 10"'
4.2 x 10"'







                                                  Entire U.S.
                                                  Entire U.S.
                                                  Entire U.S.
                                                  Entire U.S.
                                                   Urban U.S.

                                                  Urban U.S

                                                  Entire U.S.




                                                  1% of U.S.
                                                  U.S., Urban
                                                  Entire U.S.
Due to Causes Other
Than Smoking


     The data indicate  tnat  wniie the ris* of  contracting  lung
cancer is greatest  from smoking,  exposure to diesel particulate
T.ay represent a  significant  portion of  all  non-smoking-related
lung  cancer.   The  upper  limit  of  the  uncontrolled  (base)
scenario in  2000 would  represent  almost eight  individuals  in a
million, or  10 percent  of  all non-smoking-related  lung  cancer
in  the  U.S.   The  lower limit  still  represents  over  one  in a
million  individuals,  which  has  been  used  in  the  past  by
regulatory agencies as a cut-off  point  for  determining  the need
for control.   Thus,  as  indicated in  the  NPRM,  Table 4-10 shows
that  a   relatively  small but  significant  cancer  risk  may  be
attributable to  diesel  particulate exposures.   The promulgated
HDDV  controls  are  estimated  to  reduce  this'  risk  by  almost
one-half in the year 2000.

     c.     Visibility Effects

 . .  Reduced  visibility  is  one  of  the  more  readily  apparent
effects of diesel particulate.  Because  diesel  particles  are of
a diameter  most  effective in scattering light  and  their 65-80
percent  carbon  content  produces   a  high  degree   of   light
absorption, visibility  reduction  results.

     Table  4-11  presents  the  estimated visibility  impacts  of
the base- and controlled-case scenarios  in  terms of the  average
percent  reduction  due  to diesel  particulates in  1995  and 2000
urban visibility  from early  1970's levels.  As shown,  in  the
absence  of  controls,  increases  in  diesel  particulate  levels
will  result  in reduced  visibility,  ranging  from a  22  percent
reduction  in largest cities, to  4-9  percent  decreases  for less
populous  urban  areas  in  the   year  2000.   HDDV  control  is
projected  to euu these  visibility  reductions to  12  percent in
the largest  cities,  and to  2-5 percent  in  smaller  urban areas.
The   controlled-case  scenario   thus   offers  a  2-10   percent
improvement   in    visibility   over   the   base-case    scenario,
depending  on the  size  of  the city.   The lower  limit of this
impact  (i.e.,  the  effect   for   smaller  cities),  may   not  be
perceptible.  However,  the  effect for large  cities would show  a
noticeable   improvement   in   visibility.    The   promulgated
standards,   therefore,  will  provide, an  overall  benefit  that
would be most apparent  in the areas where it  was  most needed.

     d.    Soiling Effects

     In a review of  the scientific  -literature,  the DPS[5]  found
some  evidence suggesting  that because  of its  black color  and
oily  nature,  diesel particulate may  have   a   disproportionate
effect  on  soiling   conpared  to  the  effect  of  other  types of
particulate   (i.e.,   diesel  particulate   would  produce  more
soiling  than TSP  on a  gram-for-gram basis).   The black  color
may make the soiling  more apparent  to  the observer and the oily


                            Table 4-11

                  Average Reduct'ion  in Visibility
                     Due to Diesel  Particulate
          (percent  reductions from base-year visibility)

                                1995                 2000
City Size (population)    Base   Controlled    Base   Controlled

More than 1,000,000       18        12         22        12

500,000-1,000,000          7         5          9         5  .

250,000-500,000            5    "  ~  3          7         4

100,000-250,000            32          42


nature  r.ay  rake  it  Tore  difficult  to  clean.   Th=  net  effect
;.'ould  be to  increase  costs  to  the  general   public  for  rr.ore
frequent  and  more thorough  cleaning  events.   However,  because
of  th^  paucity  of  scientific   data  on the   physical  soiling
effects  of  diesel particulate  and TSP,  no  definitive statement
of these relationships can be made at this time.

     There  is  a  somewhat  larger  body of  literature available
regarding the  costs  associated with various  levels of soiling.
Summaries of  this economic literature  can  be  found in  an EPA
report   regarding   the   benefits   associated   with   diesel
particulate control,[31]  and  in  the Draft  RIA.   These  reports
conclude  that  there  are  significant  economic benefits  to be
gained  from control  of  diesel  particulates   with  respect  to

     4.    Conclusions

     Based  on  the above  projections, EPA  believes that diesel
particulate emissions are  a  serious  environmental concern with
respect  to  their impact on various  health  and welfare aspects.
It  seems apparent  that  significant  reductions   in heavy-duty
diesel  emissions  are an essential element  in  dealing with this
environmental  problem.   The  stringent  controls  on heavy-duty
diesels  and urban buses being promulgated  in the  final rule are
viewed  as  effective  means  of   reducing the  future  growth in
particulate emissions.



      1.     "National Emissions Report," National Emissions  Data
 Systerr  of  the Aeror?.etric  and  Emissions Reporting System,  U.S.

      2.     "User's  Guide  to  MOBILE2  (Mobile  Source  Emissions
 Model),"  U.S.  EPA/OAR/OMS/ECTD/TEB,  EPA-460/3-S1-006,  1981.

      3.     "Compilation  of  Air  Pollution   Emission   Factors:
 Highway Mobile Sources," U.S.  EPA, EPA-460/3-81-005,  March 1981.

      4.     "Rollback Modelling:   Basic and  Modified,"  Journal
 of  the Air  Pollution Control  Association.  DeNevers,  N.  and  J.
 Morris, Vol.  25,  No. 9,  1975.

   -   5.     "Diesel  Particulate Study,"  U.S.  EPA/OAR/OMS/ECTD/
 SDSB, October  1983.

	6.     "A  Comparative  Potency  Method  for  Cancer  Risk
 Assessment:    Application  to  Diesel   Particulate   Emissions,"
 Albert,  R.  E.,  E.  Lewtas, S.  Nesnow, T.W.  Thorsland,  and  E.
 Anderson,  submitted to Risk Analysis, 1982.

      7.     "The   Highway   Fuel    Consumption   Model:    Tenth
 Quarterly Report,"  Energy  and  Environmental  Analysis,  Inc., for
 U.S.  Department of Energy,  November 1983.

      8.     EPA    Technical    Report,   "Motor    Vehicle    NOx
 Inventories,"   Amy Brochu  and  Dale  Rothman,  EPA-AA-SDSB-85-03,
 November  1984, draft.

      9.     EPA    Technical    Report,   "Motor    Vehicle    NOx
 Inventories,"   Amy  Brochu  and Dale Rothman,  EPA-AA-SDSB-85-3,
 March 1985,  final.

      10.    EPA Technical  Report,   "Heavy-Duty Vehicle  Emission
 Conversion   Factors,   1962-1997,"   Mahlon   C.   Smith,   IV,
 EPA-AA-SDSB-84-1, August 1984.

      11.    "1982    Highway    Statistics,"    Federal    Highway
 Administration,     U.S.     Department    of     Transportation,

      12.    Argonne  National  Laboratory's  ANL-83  Projections,
 provided   to  Jim  DeMocker,  U.S.   EPA,  OAR,   as  part  of  initial
 NAPAP review,  January 1985.

      13.    "The   Highway   Fuel   Consumption   Model:    Eighth
 Quarterly Report,"  Energy  and  Environmental  Analysis,  Inc., for
 U.S.  Department of  Energy,  July 1982.


     14.   "GM    Challenges    EFA    Concern    with    Future
NOx/Particulate  Emissions,"  J.E.  Nolan  and  E.J.  Neiderbu=hl,
General Motors  Corporation, The  Environmental  Forum,  November
1984. •

     15.   1980  OBERS:    BEA  Regional  Projections,  Bureau  of
Economic Analysis, U.S. Deoartment of  Commerce,  Washington,  DC,
July 1931.

     16.   Letter  to  Mr.  T.M.  Fisher,   Director,  Automotive
Emission  Control,  General  Motors,   from  Richard  D.  Wilson,
Director,  Office  of  Mobile Sources,  OAR,  U.S.  EPA,  April  11,

     17.   Letter  to  Mr.  T.M.  Fisher,   Director,  Automotive
Emission  Control,  General  Motors,   from  Richard  D.  Wilson,
Director,  Office  of  Mobile Sources,  OAR,  U.S.  EPA,  September
25,  1984.

     18.   "Effect  of Source  Growth  on Annual  NO*  Air Quality
in Urban  Areas,"  T.Y.  Chang, Ford Motor  Company,  APCA Journal,
Vol. 32, No. 5, May 1982.

     19.   EPA    Memorandum,   ' "Off-Highway    NOx   Inventory
Development:  NEDS Methodology,"  Charles  L.  Gray, Jr., Emission
Control Technology  Division,  to  Richard  D.  Wilson,  Office of
Mobile Sources, March 5, 1985.

     20.   EPA   Memorandum,    "Stationary   Area   Source   NOx
Inventory  Development:   NEDS  Methodology,"   Dale S.  Rothman,
Emission  Control Technology Division,  to Richard D.  Wilson,
Office of Mobile Sources, March 1965.

     21.   Methodology  to  Conduct  Air  Quality  Assessments of.
National  Mobile  Source  Emission  Control  Strategies:   Fina1
Report,  EPA-450/4-80-026,  Energy  and Environmental  Analysis,
Inc.,  Arlington,  VA.    (Prepared   for   U.   S.  EPA,  Research
Triangle Park, NC), October  1980.

     22.   EPA Memorandum,  "1981-83  SMSA  Air  Quality Data  Base
for  Nitrogen  Dioxide," Richard G.  Rhoads, Monitoring  and  Data
Analysis  Division,   to  Charles  L.   Gray,   Emission  Control
Technology Division, January 11,  1985.

     23.   EPA  Memorandum,  "Comparison  of  Diesel  Particulate
and  NOx Inventories:   MOBILES  vs.  NPRM," Amy Brochu, Standards
Development  and  Support   Branch,  to  Charles   L.  Gray,   Jr.,
Emission Control Technology  Division, September  27, 1984.

     24.   "Briefing  Docur.ent"  prepared  for  EPA  Administrate r
William  D.   Ruckelshaus  by  EPA-'s  Acid  Deposition  Task  Force,
August  1,   1983,   excerpted   in   the  Environmental  Reporter,
Septeroer 2, 1983,  pp. 754-56.

     25.   "Review   of   the   National   Ambient  Air   Quality
Standards  of  Nitrogen Oxides:  Assessment of  Scientific  and
Technical  Information—OAQPS  Staff  Paper,"  EPA-450/5-82-002,
U.S. EPA, Research  Triangle Park,  NC.

     26.   "Regulatory  Impact  Analysis on  the  National  Ambient
Air Quality  Standards for  Particulate  Matter,"  U.S. EPA,  OANR,
SASD,  Research Triangle Park, February 21, 1984.

     27.   "User's  Guide  to  MOBILE3  (Mobile  Source  Emissions
Model)," U.S. EPA/OAR/OMS/ECTD/TEB, EPA-460/3-84-02,  June 1984.

     28.   Memorandum,  "Summary Emission  and Fuel  Use Forecasts
for  the   Industrial   Sector:   Base  Case  for  EPA  Emission
Reduction  Analyses,"  Craig  D.  Ebert,  ICF,  Inc.,   to  Jeannie
Austin, OPA, U.S. EPA, Washington D.C., November 12,  1982.

     29.   "Controlling   Airborne    Particles,"   Committee   on
Particulate  Control Technology, National  Academy of Sciences,
Washington, DC,  1980.

     30.   "Health   Effects  of  Air   Pollutants,"   U.S.   EPA,
Washington, DC,  June 1976.

     31.   "Health,   Soiling,   and   Visibility  Benefits   of
Alternative  Mobile  Source  Diesel  Particulate  Standards,"  Final
Report, EPA  Contract  No.  68-01-6596, Mathtech,   Inc., Princeton,
NJ, December 1983.

                                      *'  1
(Input Information:   Tables A-l through A-10)

                            Table A-l

            U.S. U r ban VMT* (pillions cf riles/year)


























     Based on MOBILE3 Fuel Consumption Model, January 21, 1985

                      Taole A-2

iModel Year
1978 ' -
T O-» A
± y t "i - -
0.590 '
n . «^sn
"of VMT
• 0.388


0. 176
0. 176
0. 176
0. 176
0. 176
0. 176
0. 176
Used   in   MOBILES  Fuel   Consumption   Model   to   convert
nationwide VMT into urban VMT;  based on MOBILES conversion
factor analysis.

Table A-3
Urban Fraction jf VHT
Model Year
- 0.690

Used   in   MOBILES   Fuel   Consumption  Model   to   convert
nationwide VMT into urban VMT; based  on  MOBILES conversion
factor analysis.

                            Taole A-4
                Liant-Dutv Diese-l Sales Fractions
                     "Best  Estinace"
                             "Low Grcwth"*
Model .Year

• 1975




 . 000
      "Low  Growth"   fractions   used   in  diesel   penetration
      sensitivity analysis.

                      Table A-5

          Heavy-Duty Diesel  Sales  Fractions*
Heavy-Duty Truck Class
Model Vear
. .280
.660 •
- .886
" 1.000
- -.920
' .721
Based on MOBILE3 conversion factor analysis.

                        Table A-6
                Fleet-Averase Heavy-Duty
        Emission Conversion Factors ( BHP-hr/rrd ) *
Model. Year
. 1971
, --1994 —
' Gasoline

Based on MOBILES conversion factor analysis.

                           Table A-7

     Growth Rates  and Assurr.ptions Used in FKM HOx Analysis
VMT Growth Rates*
(%/year, compound)
Vehicle 1982-1995 1982-2000
Class Urban Nationwide Urban
LDV*« +1.9
LOT** +2 .2
HDGV +0.4
HDDV + 4.7
Source Category _
__ Stationary Point***
Stationary Area
• — • —
+1.9 +1.9
+2.2 +2.1
0.0 +0.6
+3.2 +4.2
Source Assumptions
Growth Rate
+ 1.3
+ 0.8
+ 1.9
+ 0.2
+ 2.9
* *
     Based  on  MOBILE3   Fuel  Consumption  Model,  urban  growth
     rates also used in diesel particulate analysis.
     Light-duty  urban  fractions of  VMT  are  assumed to  remain
     constant with  model year;  therefore,  urban and nationwide
     growth rates are equal.
***  Stationary  point  source  growth   rate  assumes  a  certain
     level  of   future   NOx  control,   based   on   ICF,   Inc.,
     projections;  point  source  emissions   included   only  in
     nationwide NOx projections.

                               T=ble A-8

          SMSAs Modelled for NO."  (Design Values,  pprr  NO,)
H?RM Analysis -

Boston    (0.050)

Chicago    (0.060)

Cleveland    (0.048)

Nashville    (0.047)
                  Interim Analysis -,-

                  Boston   (0.036)

                  Chicago   (0.052)

                  Nashville   (0.053)

                  New York   (0.036)
 Philadelphia    (0.046) Newark    (0.045)

.Steubenville    (0.040) Philadelphia    (0.039)

 •Denver    (0.046)      Seattle     (0.048)

 •Reno    (0.048)        Tucson    (0.037)

                       Wash., D.C.    (0.036)

                       •Denver    (0.041)

                       •Reno    (0.043)
FRM Analysis -

Chicago   (0.044)

Cincinnati   (0.036)

Nashville   (0.053)

New York   (0.037)

Newark   (0.040)

Philadelphia  (0.040)

Pittsburgh   (0.035)

Wash., D.C.   (0.037)

•Denver   (0.052)

•Reno   (0.043)

Los Angeles   (0.059)

Anaheim   (0.045)

Riverside   (0.042)
California  SMSAs  included  only  in  the   FRM  analysis;  future
projections based on CARS data.
NOZ concentrations at or above 0.040 ppm (75% of std.)
N02 concentrations at or above 0.035 ppm (66% of std.)
Interim  analysis  results   presented  in  Motor   Vehicle  NOx
Inventories  (Technical  Report),  and  letters  to  T.  M.  Fisher
(GM) and  Donald R.  Buist  (Ford),  all  contained in  the  Public
High-altitude SMSAs.

                            Table  A-9
        Low Altitude  NOx  Emssion  Rates  and Assumptions
                Different  Tnan  MOBILES  Values  for
           Emission  Inventory and Air Quality  Analysis
Base Case:
Vehicle Model
Emission Rate Useful
Type Year ZM[1] DR[2] SEA[3] Life[4]
LDGT 1987 +
LDDT 1978-80
1985 +
HDGV 1987
HDDV 1987
Cl. II a[6] 1987
1988 +
LDDT 1978-87
' Cl. I[5] 1988+
Cl. II a[6] 1988+
HDGV 1987
1997+ c
HDDV 1987
1997 +




«» v
— —
— —
— -

as Base
as Base
as Base

- Full
 [1]  Zero-mile emissions  (g/mi).
 [2]  Deterioration  rate  (g/mi-lOK mi).
 [3]  Selective Enforcement Audit.
 [4]  Certification  to  half or  full useful  life.
 [5]  Le'ss than 6,000  Ibs  GVW.
 [6]  6,001  -  8,500  Ibs GVW.

                           Table A-10

                High-Altitude tiOx Emission Races
          and Assumptions  Different  than !:OBILE3  Values
         for Emission Inventory  and  Air Quality Analysis
Base Case:



    Emission Rate     Useful
ZMC11   DR[2]  SEA[3)  Li£e[4]
   Same as Low Altitude

   Same as Low Altitude
1997 +
                         Same as Low Altitude
• 6.0/5.0)


   Same as Low Altitude

   Same as Low Altitude
1997 +
Same as Base Case
3.65 0.10 40%
3.62 0.10 40%
2.97 0.09 -40%
2.89 0.09 40%
2.83 0.09 40%

                         Same as Low Altitude
 [1]  Zero-mile emissions  (g/mi).
 [2]  Deterioration  rate  (g/mi-lOK mi).
 [3]  Selective Enforcement Audit.
 [4]  Certification  to  half or  full useful  life.

                           CHAPTER  5

                       COST EFFECTIVENESS

     The  cost effectiveness of an  action  is the measure of  its
 relative   economic   efficiency  toward  achieving  a   specified
 goal.   It is primarily useful in comparing  alternative means  of
 achieving  that  goal.   The  cost  effectiveness  of  the  final
 particulate and NOx  standards  analyzed in  this  report will  be
 the  subject  of this  chapter.   Before  the  final  analysis,  an
 overview   of  the  cost-effectiveness  analysis  in   the  Draft
 Regulatory Impact Analysis (RIA) and  a summary and analysis  of
 the  comments received will be presented.

 I.   Overview of  NPP-4 Analysis                                 _

     In the Draft RIA, EPA determined the  cost  effectiveness  of
-the  proposed  standards  in  terms  of the dollar  cost per ton  of
 particulate  or  NOx  emissions  controlled.   These  values  were
 used to  make  comparisons  with the  cost  effectiveness of  other
 mobile  and non-mobile source  control strategies.

     To determine cost effectiveness, two  pieces of  information
 were necessary:   the  costs   and   emissions reductions  of  the
 strategies to be examined.   The  costs and  emissions  reductions
 used were  those  associated  with  an average  vehicle  on  a
 per-vehicle basis,  rather  than the  total   costs and  reductions
 for  the entire fleet.

     The  costs .used were   those  determined  in  the  economic
 analysis  cf the  proposed  standards.   The emission  reductions
 were  calculated   for  each  year   of   the   vehicle's  life  by
 multiplying the  vehicle's  miles travelled  (V1T)  by  an  average
 per-mile  emission  reduction.   The  annual  VMT values  used  were
 those  determined by Energy and Environmental Analysis  adjusted
 to  reflect  EPA's  lifetime  estimates.   For  heavy-duty  diesel
 vehicles, a composite V4T was calculated by sales  weighting the
 individual  values   for   light,   medium,   and heavy  heavy-duty
 diesel  vehicles  (LHDDV, MHDDV,  HHDDV).   The average  per-mile
 emission  reductions  used  were  developed- using  information  from
 the  MOBILE2.5  emission  factor  model and the  Diesel  Particulate

      Two    approaches   were    used   in    calculating    the
 cost-effectiveness   values  for  the  -proposed   standards:    an
 annual   approach   and a  lifetime   approach.   With  the  annual
 approach,  costs  were  allocated:   1)  to  each year  in  which
 emission  reductions  were  produced,  and 2)  in proportion to the
 size   of   these   annual   reductions.    The  result   was   a
 cost-effectiveness  value  which is  applicable  at  any point  in
 the  life  of the  vehicle, as well  as over  the  vehicle's entire


lifetime.    This  approach   allowed   for   comparisons  on   a
consistent  basis with  recent EPA  cost-effectiveness  estimates
for other  mobile and stationary source particulate control, and
stationary source NOx control.

     With  the   lifetime   approach,  the  lifetime  costs  were
discounted  to  the year of vehicle  purchase  and  then divided by
the  undiscounted  total  lifetime  emissions  reductions.   The
lifetime   approach   was   only  used  in  conjunction  with  the
proposed  NOx standards  to allow  comparisons  with  past  mobile
source  cost-effectiveness studies,  where  only this  method was

     Special  considerations   in  the case of  particulate  matter
led     to     the    determination    of    several    different
cost-effectiveness values for each standard.  Since the effects
of  particulate  matter are highly  dependent  upon  particle size,
emissions   reductions   and   cost-effectiveness   values   were
determined  on  a total,  inhalable, and fine basis.*  Also, since
the  great  majority  of  people   who  are   exposed  to  NAAQS
violations   for  particulate   matter  live   in   urban  areas,
emissions   reductions   and   cost-effectiveness   values   were
determined  on both  an  urban and  a nationwide basis.   For the
urban   estimate  the  only   change  made  was  that  emissions
reductions  in  non-urban  areas  were  excluded? no  changes  were
made in the  cost estimates.

II.  Summary and Analysis of  Comments     .     _	 ._

     There   were  very   few   comments  received   that   dealt
specifically   with   the   cost-effectiveness  methodology   and
procedures  used  in  the  Draft  RIA.   Comments received  on the
cost-effectiveness  estimates  that  primarily  address either the
costs  of  control,  or the  emissions  reductions  obtained,  have
been   reviewed   in   the   respective  chapters  on  Economic  and
Environmental  Impact.

     There  remained  only  three  comments   specific  to  cost
effectiveness.   The  Department of  Energy  (DOE)  presented its
own  cost-effectiveness   estimates   that  indicated  that  EPA' s
estimates  may be somewhat low.  In their methodology, the costs
used were the  undiscounted costs  of control,  and the standards
considered differ  slightly  from  those  that EPA  considered.
      Total   particulate  is   all   suspended  particulate  matter
      regardless    of   diameter,   inhalable   particulate   is
      considered   to  be  all  particulate  matter  less  than  10
      micrometers   in   diameter,   and   fine    particulate   is
      considered   to  be   all  particulate  matter less  than  2.5
      micrometers  in  diameter.


Several environmental  groups  called attention to  the  fact that
the cost-3ffectiveness  estinates  for the  0.25 g/BHP-hr  and the
0.1 g/BHP-hr  standard  for urban HDDEs  in 1990 were equivalent.
Based  upon  this,   they  questioned  EPA1 s  choice  of  the  more
lenient  level of  control.    Finally,  DOE  took  issue with the
alleged use  of a  100   percent  discount  rate  for  NOx  enissions
from  elevated  stationary  sources.   In  their  opinion,  this
effectively   renders   any   comparison   of   cost-effectiveness
between mobile and stationary sources meaningless.

     DOE' s   practice   of   using   undiscounted   costs  appears
inappropriate  to EPA,  considering  the basic economic concept of
the  time  value of  money.  Since  DOE did not present its cost
estimates  in a   detailed   fashion,  it  is  not  possible  to
determine  how  much  of  the  difference  in  cost-effectiveness
"values  can   be  explained  by  this  difference  in  accounting
methods.   In any  case,  even   if  the  basic   technological
economic,  and environmental concepts were  the  same,  the use of
undiscounted   costs  will  lead  to  higher  cost-effectiveness
values.   Since  there  is  no apparent   reason  for using this
approach,  it  will not be  considered  further.

     In the  Draft  RIA, EPA did estimate the same value for the
cost  effectiveness  of  a  0.25  g/BHP-hr  and  a  0.1 g/BHP-hr
particulate standard for  urban  HDDEs in  1990.  However, EPA did
indicate  that it believed  that in  fact the more  stringent 0.1
g/BHP-hr  standard  would  actually be  less  cost  effective, for
several   reasons.     The    maximum  benefit   and   least   cost
applications  would  have  already  been  used to  meet  the 0.25
g/BHP-hr  standard  (with  averaging), so  that  subsequent  use of
traps   on  additional   engines   might   be   somewhat   less  cost
effective.  Other  factors cited which argued for higher cost at
the  0.10  g/BHP-hr  level are greater development costs, the need
to  design  to  lower  low mileage-target   emission levels, the use
of  higher  quality  components,   the   probable   need  for  more
frequent  trap regeneration,  and  the increased risks  associated
with in-use compliance.

     As  will  be seen   in the updated analyses here  and  in the
Alternatives  Chapter,  the cost  effectiveness of a  0.10 g/BHP-hr
standard  does turn out to be  somewhat worse than that of  a 0.25
g/BHP-hr  standard,  confirming EPA1 s original position.  It must
also  be noted that  cost  effectiveness  is only  one factor used
by   EPA  in   deciding   between  control   options;   technological
feasibility  has been   the  primary  basis for the  decisions in
this  rulemaking because  of  the  statutory provisions governing
both  the  NOx and particulate  standards.   It was on the basis of
technological constraints  that   EPA  decided   against  a 0.10
g/BHP-hr  standard  for  1990.


     Contrary  to  the assertion  by DOE,  EPA did not discount the
emissions-  reductions from stationary sources of NOx anywhere in
its   cost-effectiveness   analysis.    While   some  degree   of
discounting    emissions    reductions    based    upon    spatial
considerations   may  be   appropriate   in   comparing  the  cost
effectiveness  on  an urban basis,  where  stationary sources have
relatively  little impact  upon  breathing  zone concentrations of
NO2»   this  would  be   less  appropriate   for  regional  scale
considerations.   Therefore,  this analysis  has  not discounted
the  NOx   emissions   reductions   from  stationary  sources  when
making comparisons  of  cost effectiveness.

Ill. Updated Cost Effectiveness  Analysis

     A.    Changes  in  Analysis

—	There has been no  basic change in the methodology used to
determine  cost  effectiveness.   The  bases  for  determining the
costs   and    emissions    reductions   to    be   used    in   the
cost-effectiveness  analysis remain  the  same,  with their values
changing only  so  much as the  estimates have  been  improved.

     In  the Draft  RIA,   the  differences  between annualized and
lifetime   cost   effectiveness    were   explained   thoroughly.
Mathematically,  the difference  lies solely  in how the benefits,.
i.e.,   emissions   reductions,   are  handled.    Lifetime  cost
effectiveness   reflects   the   case  in   which  the  emissions
reductions  are  undiscounted;   annualized  cost  effectiveness
reflects   the   case  in   which   the  emissions   reductions  are
discounted at  the same rates  as  the  costs.

     Discounting    emissions    reductions    assumes   that   the
emissions  reductions are  worth more at the  present  time than in
the future.   For NOx, where exceedances of  the  ambient  standard
for NO2  are  projected   in  1995  and  2000,  but  not presently,
emissions  reductions  may actually be worth  more  in the  future
than  they  v/ould be  now.   In the  case of particulate matter, for
which  many  areas  of the country  already  exceed  the  ambient
standard,  this is not the case, and it could be argued  that the
sooner  reductions  are  obtained  the better.   Thus, it is not
clear  to  EPA  if,  or how much emissions  reductions  should be
discounted  over   time.    Therefore,   the  estimates   of   cost
effectiveness   in the  final  analysis- are  shown  using   several
different  discount   rates  for the  emissions  reductions.  The use
of  various discount  rates here  allows  for proper  comparisons of
cost   effectiveness  to   other   mobile  and  stationary  source
controls   to  be  made,   and  for  the sensitivity  of  the  cost
effectiveness  to the discount rates  to  be  established.

      In the  Draft  RIA,  the  costs  and  emissions reductions
estimates  for the  later  year,  1990,  standards  were  presented as
 incremental  to the  values for  the 1987  standards.   This yielded
an incremental,  or marginal cost effectiveness.   In the  final


analysis of this  chapter,  this  has also been  done  for the 1991
and  1994 • standards,  and  is referred  to  as  the marginal  cost
effectiveness.*  In addition, cost-effectiveness  values for the
combined standards  have also  been determined for  these  later
year  standards,  and   are  referred  to   as   the   total  cost

     Updated estimates  of  the  expected annual  and  lifetime per
average vehicle VMT  are shown  in Table 5-1.   The VMT for HDGVs
has  not changed  since  the Draft  RIA,  and  the estimates  for
line-haul  truck  VMT are  the same  as for  HHDDVs in  the  Draft
RIA.   The   estimates  shown  for  non-line-haul  trucks  for  the
various  model  years  were  determined by  taking weighted  sums
across  the VMTs  for  LHDDVs and MHDDVs as  given in  the  Draft
RIA.    The  weightings  were   derived  from   projected  sales
fractions  in  each year by  class,  and corresponding projected
diesel  sales fractions.[1]  As  these  change  over time, so does
the  average VMT  for  non-line-hauls as a  whole.  For LDT^ and
LDT2,  separate estimates  of VMT were derived,  which  was  not
done in the Draft RIA.   These were derived by taking  the annual
average  mileage  accumulation rates  in MOBILE3  and  multiplying
each  year's  VMT  by   a  survival  fraction  derived   from  the
registration data  in MOBILES.[2]   The VMT for  urban buses has
been updated to reflect more recent EPA data.Cl]

     The emission rates  for the  proposed standards  vary over
the  life  of the  vehicle  and are summarized  in  Table 5-2.  The
particulate  equations  were derived   using  the  methodology  as
described  in   the  Diesel  Particulate Study.[4]   For  NOx,  the
values  are derived  from the MOBILES  emission factor  model, and
represent  the  actual in-use emissions including misfueling and
tampering.[2]   The  slight  difference  in  form  for   particulate
and  NOx reflects  the differences  in  how  the emission  rates for
these  two  different  pollutants  are determined.   Note here that
varying  emission  rates for  each  year  of  the vehicle's  life are
being used in  this  analysis; in the earlier analysis  an average
rate  determined  at  the  vehicle's half   life  was  used.   This
change  leads to improvements in  the accuracy of  the  estimates.

     3.     Results of Updated Analysis

     The  emission reductions  and  cost-effectiveness  estimates
for  the NOx and  particulate standards are  shown in  Tables 5-3
and  5-4,   along   with  the  costs  from  the   Economic  Impact
chapter.   These  costs  represent the  net  present value in the
year  of  sale   to the  consumer,  using  a  10  percent  discount
rate.   It  includes  both the first price increase and increased
      1991  standards  marginal  from  the  1988  standards,   1994
      standard marginal  from the 1991  standard.

                                        Table 5-1
Annual and Lifetime Per Average Vehicle \MT (miles)Cl]

                   HDDS Nbn-Line-HaulC2,3]
- 3
"~" 4,986
- 5,058
- 980
3,790 '
Total    123,648
 127,691   110,190  184,363  183,418  184,048  540,000   527,740
 LIJ  Urban fraction of travel for HDEVs:
     Non-line hauls = .475,  line hauls =  .176, buses  =  1.000, all =  .288
 £23  Changes  in  \MT by  model  year  due to  changes  in  relative  total  sales
     fractions  of heavy-duty classes IIB-V and VI-VIIA.
 [3]  HDDE all  classes \MT  can be approximated by taking a weighted average of
     non-line  hauls  and  line  hauls.   Relative  weights  are  .635 and  .365,

Base Case:
(no further
1988 Standard
0.60 g/BHP-hr
1991 Standard
0.25 g/BHP-hr
0.10 g/BHP-hr
for urban bus
1994 standard
0.10 g/BHP-hr
Table 5-2
Annual Per Mile Emission Rates
(grans/mi le)

Vehicle Type


Urban Bus



Urban Bus



Urban Bus

Urban Bus
. .0000
[1]Emission  rates  vary  slightly  with  model  year   due   to
     changes  in  the conversion  factors  between  g/BHP-hr  and
[2]  Zero-mile emissions  (g/mi)
[3]  Deterioration rate  (g/mi x (years-0.5))


                       Table 5-2, Conf d

                 Annual Per Mile Emission
                         (orams'/mi le)
Base Case:
2.3 g/mi LOT
10.7 g/BHP-hr HDE
1988 Standard
1.2/1.7 g/mi

6.0 g/BHP-hr HDE
1991 Standard
5.0 g/BHP-hr HDE

Vehicle Type





1991 .

' 0.94
- 4.67

• "• -.0000
[1]  Emission  rates  vary  slightly  with  model  year  due  to
     changes  in  the  conversion  factors between  g/BHP-hr  and
     g/mi .
[2]  Zero-mile emissions (g/mi)
[3]  Deterioration rate (g/mi x 1000 mi)

                                             Table 5-3

                               Urban Particulate Cost Effectiveness
    Option  .

1988: 6.60

1988: 0.60
1991: 0.25
for urban buses

Bus Only 1991
1988: 0.60
1991: 0.25
for urban buses
1994: 0.10
Benefits (tons) Rate
Costs($) (1]
(625-728) (2)
Cost Effectiveness ($/ton)
Discount Rates for Benefits
[1J   Costs represent  net  present  value  in year  of sale  of the  total  cost to  consumer,  using  a
     constant 10 percent discount rate.
(2)   Figures in parentheses indicate marginal values from  standard levels of 0.60 g/BHP-hr in 1988
     and 0.25 g/BHP-hr in 1991 (0.10 g/BHP-hr for urban buses).
(3|   Cost, benefits, and cost effectiveness for urban buses only.
14]   This value  calculated  by taking  a vreighted  average  of the  heavy-duty bus  and  non-bus trap
     equipped costs to go from no control to .0.10 g/BHP-hr in 1994.

                                        Table 5-4

                                 NOx Cost  Effectiveness
Option (g/BHP-hr)   Costs($)(1)
LOT: 1.2/1.7, 1988

HDC: 6.0, 1988

HDE: 6.0, 1988
     5.0, 1991





                                   Benefits  (tons)  Rate







 .093    .079
(.073)  (.061)
3.863   3.126   2.634
(.726)  (.583)  (.488)
                                          Cost  Effectiveness ($/t'on)
                                         Discount Rates for Benefits
- 44-99
[1]   Costs represent  net  present  value  in year  of sale of  the total cost  to consumer,
     using a constant 10 percent discount  rate.
(2)   Figures in parenthesis  indicate  marginal values from standard  level  of  6.0 g/BIIP-hr
     in 1988 for HDEs.


operating  coses.    Where  applicable,   the  total  and  marginal
values  fc?r  each standard  level have  been  presented.   For  the
particulate  standards,   only  the  urban  cost  effectiveness  is
given  in  this  analysis  as the  nationwide value  has  not  been
used in comparisons with other sources.

     The  costs and cost-effectiveness  values  presented  here
represent the  long-term  values  for  each of  the standards as the
fleet  stabilizes  in its response  to  the change  in  standards.
In the  short  term,  the  costs  associated with  the standards will
be  somewhat   higher   as  discussed   in  the   Economic  Impact
chapter.     This    would     in    turn    result   in    higher
cost-effectiveness estimates in the short terra.

     The discount  rate  used for the benefits  can  have  a marked
effect  on  the benefits and cost-effectiveness  values.   As seen
in  the tables,  this  results in  a 40-60  percent increase  in
cost-effectiveness   values    in   comparing    results   using
undiscounted  benefits  and  those discounted at  10  percent.  The
cost estimates  from Chapter 3 used a  10 percent discount rate.
Thus,  the  cost effectiveness  estimates  at  a  0 percent discount
rate  are  equivalent to  the  lifetime  cost-effectiveness values
as  described  in   the  Draft  RIA,   and  those  at  a  10 percent
discount  rate are  equivalent to  annualized  cost-effectiveness

     C.    Comparison to Other Control  Strategies

     1.    Particulate

     Table 5-5  presents  an update  of  Table  6-5 in the  Draft RIA
comparing  the  relative  economic   efficiencies  of   controlling
particulate   emissions   from  other   mobile   and   stationary
sources.   Other than  updating  the values  from  1983  to 1985
dollars, based  upon the consumer's price index for new cars and
the  producer's  price  index   for   industrial   commodities  (5.9
percent and  2.4 percent respectively)  no changes have  been made
in   the  estimates  for   other  sources,  taken  from  the  DPS
report.[5]  As  in~the  Draft RIA, the comparison is presented on
the  basis of  total,  inhalable, and  fine particulate; and the
stationary  values  have  been  adjusted  to reflect  the   relative
breathing  zone  air quality impact of those emission  compared to
that of diesel  emissions.£3,43

     The  updated estimates of the cost-effectiveness values for
particulate  control for HDDV are  generally  equivalant  to those
in  the Draft RIA.  Therfore, as  would be expected,  the  figures
in  Table  .5-5 suggest that HDDV controls remain quite  favorable
when  compared to  stationary  source  controls,  regardless of the
size  of particulate examined.   Only  the control  of  wet  cement
kilns  appears to  be significantly  more cost  effective than any
of  the HDDV standards.  Thus,  it  is a fair conclusion to state


                           Table 5-5

              Annual Cost Effectiveness  Comparison
           for Particulate Control  Of Urban HDDVs  and
     •Other Mobile and Stationary Sources  (S/ton)[1,2,3,4]

Sources [5]
Cement Kiln
Bus Only 1991 Standard
MODE 1988 Standard
LDDT (.26,1987)
HDDE 1991 Standard
HDDE 1994 Standard
LDDV (.2,1987)
Kraft Smelt Tank
Electric Arc Furnace
Borax Fusing Furnace
Industrial Boiler
Kraft Recovery Furnace
Lime Kiln (baghouse)
Electric Utility
Lime Kiln (ESP)
iculate Size
Basis [6]
[1]   Stationary   sources   are  discounted   to  reflect   their
     relative ground level effect.
[2]   1985 dollars.
[3]   Emissions reductions discounted 10 percent.
[4]   For simplification,  the midpoint of the  ranges  were  used,
     where applicable.
[5]   Ranking based upon inhalable particulate values.
[6]   See References  3  and  4  for other  mobile and  stationary


that  the  HDDV particulate  standards are  quite  cost  effective
when  compared to  stationary  source  and   other  mobile  source

     2.    NOx

     Tables 5-6  and  5-7 present  updates of Tables  6-8  and 6-9
in  the  Draft  RIA  comparing the relative  economic efficiencies
of  controlling NOx emissions  from various  mobile and stationary
sources.   The  values  for the  more stringent  standard  for LDVs
has  been  updated  from  1984  to  1985 dollars  by  2.4  percent,
based  upon the  consumer's  price  index  for new  cars,  but are
unchanged   otherwise.[5]    The   values   associated   with  I/M
programs  for  LDVs  represent more recent EPA estimates.[10]  The
cost  effectiveness values associated with  the stationary source
controls  of  NOx  have  been  updated  to   reflect  more  recent
analysis  performed by EPA1 s Office  of  Air Quality Planning and
Standards,  the South  Coast  Air Quality Management District, and
EPA Region  IX. [6,7,8,9]

     As with  diesel particulate,  the updated estimates for cost
effectiveness of  the  NOx standards  are  generally equivalent in
the updated analysis  compared to  the  results  in the Draft RIA.
The estimates  for  HDGEs  have  increased from $15 to $278/ton and
from   $55  to   $l5l/ton*   for   the   early  and   later  year
standards.[3]   This   reflects  changes  in  the  emission  factors
from MOBILE2.5 to  MOBILES  and increases in the costs associated
with  the HDGV NOx standards.   Since the  emissions reductions
and  costs  associated  with  the  HDGE  NOx  standards  are  small f
even  slight changes   in  their  estimates can have large  effects
on  the  cost effectiveness as has been seen  to be  the case.

      The  final  NOx  standards  for  LDTs  and HDEs  remain  quite
favorable  in  cost-effectiveness  comparisons to other mobile and
stationary  source  controls  of NOx.  The final NOx standards for
LDTs  and  HDEs have lower cost-effectiveness values than  almost
all  of the other  mobile or stationary  source control  options.
If  the  stationary  source  NOx  emissions  were  discounted  to
reflect  their relative  ground  level effect,  as was  done for
particulate,  the cost effectiveness of the proposed LDT  and HDE
NOx standard would compare  even more favorably.
     Using undiscounted benefits.


                            Table 5-6

                Lifetime Effectiveness  Comparison
                of NOx Control for Mobile  Sources

                                              Cost Effectiveness
	SourceU]	      ($/ton) [2,3]

HDDE 1988 Standard                                     22
HDDE 1991 Standard                                  35-81
HDGE 1991 Standard                                    183
LOT 1988 Standard                                     263
HDGE 1988 Standard                            .       278
LDVs (I/M, where presently exists  for HC/CO)          527[4]
LDVs (I/M, where none  presently  for HC/CO)           2290[4]
LDVs (1.09 to 0.4 g/mi)   -                        .  . 2460[5]
[1]  Ranked according to midpoint of range.
[2] -1985 dollars.                        -.      -        .  .T._
[3]  Emissions  reductions undiscounted.
[4]  See Reference  10.
[5]  "Cost  Effectiveness  of  Large  Aircraft  Engine  Emission
     Controls   -  Final  Report,"  U.S.  EPA,  OAR,  OMS,  ECTD,
     December 1979.

                           Table 5-7

             Annual Cost-Effectiveness Comparisons
    for NO* Control of LDTs and HDEs and Stationary Sources
HDDE 1988 Standard
HDDE 1991 Standard
Industrial Residual Oil Boilers
HDGE 1991 Standard
LDT 1988 Standard
HDGE 1988 Standard
Industrial Coal Boilers
Internal Combustion Engines
Cement Kilns (Calif.)
Stationary Gas Turbine
Internal Combustion Engines (Calif.)
Glass Melting Furnaces (Calif.)
Refinery Heaters and Boilers (Calif.)
Cost Effectiveness

      •   162C4]
L1JRanked according to midpoint of range
[2]  1985 dollars
[3]  Emissions reductions discounted 10%
[4]  See References 6 and 7
[5]  For applications  in  Southern  California, see  Reference 8
     and 9


     D.    Conclusion

     The   cost   effectiveness  of   the  particulate   and   NOx
standards  is  favorable when  compared with other  mobile  source
control strategies.   This  is  also  true when these standards are
compared  with  stationary  sources.   Therefore,  based on  this
above  analysis,  the  standards  appear to  be  a  cost-effective
means  of  reducing  particulate  and  NOx  emissions  compared  to
controlling these pollutants  from other sources.

             ' Appendix A
Summary and Analysis of Comments on the
Proposed Particulate Test Procedure for
       Heavy-Duty Diesel Engines

             Surjtiary and Analysis of Comments on the
             Proposed Particulate Test Procedure for
                    Heavy-Duty Diesel Engines

      Following   the   publication   of   the   NPRM,    the    HDD
 manufacturers   submitted  written   comments   on  the   proposed
 particulate  test procedure.   Also,  a meeting  was  held  between
 the  Engine  Manufacturers  Association (EMA)  and EPA on  January
 28,  1985  during which  HDD  particulate  test procedure  details
 were   discussed.   A  memorandum  describing   this   meeting   is
 available   in   Docket  A-80-18.   The  written  test  procedure
 comments  as well as  the  verbal comments  made at this  meeting
 are summarized and  analyzed  below in four groups.

      The  first  group  includes those which  were well  supported
 by data  or  engineering  analysis  and   which  will  not  affect
 measured   particulate  mass.    The   recommendation   here  is   to
 essentially  accept  the  test  procedure  revisions  contained  in
 these comments.

—  - -  The  second  group  of  issues  include those which  were  not
 well  supported  by  available  data  or engineering  analysis  and
 where  the  available  data   indicated  that  the   change  could
 significantly     affect    measured    particulate   mass.     The
 recommendation  here  is  to  deny   these  requests   for   test
 procedure  changes,  until  it  becomes  clear  that  such  changes
 will  not  affect particulate  measurements.

      The   third  group  of  issues   are   those  upon which  -EPA
 requested  comment  in the  proposed rule,  and  the fourth  group
 are   those  which  do  not  relate  to  Subpart  N but  are  still
 related to heavy-duty engine testing.

      The  analysis of each issue  begins  with a short  description
 of the aspect of the  test procedure in   question.   The comments
 made   on  this   aspect  are  then  summarized.    Finally,   the
 available  information  relating  to  the  issue  is analyzed  and a
 recommendation is made.

 I.   Recommendations Accepted  by EPA

      Exhaust System Len-gth

      Section   86.1327-87(f)    of   the    proposed   regulations
 specifies  that  the distance  from  the  manifold to  the end  of
 chassis-type  exhaust  system  should be  a maximum  of   12  feet.
 Also,  the  length of  exhaust   system  tubing  from exit  of  the
 chassis-type system or  from the  manifold to the dilution tunnel
 shall be no more than 12  feet (maximum) , if  uninsulated,  or  20
 feet   (maximum),  if  insulated.   This  tubing  shall  be  made  of
 stainless  steel.

      Summary of  Comments:  Ford  is  concerned that:   1)  12  feet
 of chassis-type  system  may  be too  short for all in-use systems,
 and  2)  two  maximum   exhaust   system   lengths  are  possible,


depending  on whether a chassis type  system  is  used (32 feet  is
maximum) or  if not  (20 feet .

     EMA expressed  concerned about  the  following three  issues:

     1.    "EPA  has   addressed   the  issue  of  exhaust   system
design  in  the existing Final  Rule  for  gaseous  emissions  (48  FR
52227)  considering  the effect of  upcoming particulate control.
In  Section 86.1327-84(f)(2)(i) of  this final  rule, EPA permits
a total of 32 feet  length from engine to  tunnel inlet."

     "Engine  manufacturers  have  all  completed  permanent  test
cell  installations  following  these  guidelines.    EPA  has  made
some significant changes  in  the  current  proposed  rule (49FR  <§
40314)  that  will  cause   significant  modifications  and   undue
expense.    EPA  states   that  both   a   chassis-type   and   a
facility-type exhaust  system may  be used.   It  is not clear that
they infer "simultaneously."   If  EPA intends to permit only  one
or  the  other  system, then  the   individual  lengths   permitted
would  require major test cell  modifications  to  most facilities."

     2.    EMA  is   also  concerned  that  the material  that  was
specified  for the  tubing  is  stainless  steel which  they believe
(a)  is different from the gaseous  emissions  rule, and   (b)  is
not necessary.

     3.    EMA  also requested that the rules exclude  insulation
in  the  vicinity  of  instrumentation  such as smokemeters.

     Mack  also expressed concern  on the issue of exhaust  tubing
lengths.   Their  position,  while raised separately,  is  generally
.the same as  the  EMA position.

     Analysis  of Comments  and Recommendation:   The wording  in
the  proposed test  procedure  regarding  allowable exhaust  system
lengths is somewhat  ambiguous.   It  was   intended  to  specify  a
total  exhaust  system  length  of  32 feet,  with the  option  of
using   either   a chassis   type   system  (with   its own   length
limitation),  a  facility type  system or  both  together.

     The  final   rule  limits  the  amount  of uninsulated  tubing to
12  feet, which  limits  the  amount  of conductive cooling that  can
be  achieved from   the  tubing   walls  at  a   place where  the
temperature  differential   is   greatest.   Yet,   having  up   to  12
feet  of  uninsulated  pipe  provides  reasonable flexibility  for
engine changes  without the incumberance  of  insulation.   If  the
typical length  of   an  engines  chassis exhaust  is  greater  than 12
feet,  use  of the typical  length  is permitted,  but   only 12 feet
of  it  can  be uninsulated.

     A provision  should  also  be  made  for  up  to   18  inches of
uninsulated  tubing  for instrumentation  (an  in  line smoke  meter,


 for  example)  since  such  instrumentation  is  required  by  E?A.
 However,  to maintain a consistent  limit  on uninsulated  tubing,
 such   an   uninsulated  portion  should  be  counted  towards   the
 maximum total  uninsulated length  of  12  feet.

      Based on  EPA's experience,  it appears  that the  type  of
 tubing  steel   should  be   irrelevant  for   diesel   particulate
 testing since  it  is soon  covered  with a  layer  of  particulate
 and  further  wall  contact of  the  exhaust  stream is unlikely.
 The  only  exception  would  be  steel  with  an  extremely rough
 surface   which   persisted  despite   a   layec   of  deposited
 particulate,  which  could  occur if  a  rustable steel were used.
 This   could cause  additional   deposition.    Thus,   the   tubing
 specification   should  be  changed   to   include  typical   in-use
-exhaust system materials,  which could  reduce  costs  for  some
 laboratories.   However, the steel should  be  free  from any rust.

 Thus,   in  summary,  it  is  recommended  that   the  exhaust  system
-specifications be changed  and  clarified to  include  provisions
 for  1) a  total length of  32  feet, 2)  a system  which  can  be
 either chassis  or  facility  type,   3)  no more  than 12  feet  of
 uninsulated tubing,  4)  tubing  in   vicinity  of  instrumentation
 can  be uninsulated,  and  5)  tubing  can  be made of typical in-use
 materials, but must be free of  rust.

      Dilution  Air Filtering or  Backpressure  Measurement

      Section  86.1310-87(b)(1)(iv)(B)  of   the   proposed   test
 procedures requires  the primary and  secondary dilution  air  to
 be filtered if background particulate is  not measured.

      Summary of Comments:  EMA commented that  if  a  manufacturer
 does   not   filter  dilution  air  or  measure  and  correct   for
 background  particulate,    the  manufacturer   will   only    be
 penalizing  itself  and  not  the  environment  (i.e.,  this  will
 cause  a  higher particulate  emission  calculation).   Thus,   EMA
 recommended that  the  engine  manufacturer should  be given  the
 option to  simply  use  good  engineering judgment  to  account  for
 background particulate (i.e.,  filter  dilution air, measure  of
 background  particulate  levels,  or  ensure  backpressure  levels
 are  sufficiently  low so as to be ignored).

      Ford  also  believed  that  need  for  filtering or  background
 correlation  should  be  established by  the  manufacturer.  .  It
 recommended  monthly  background  checks,   and   if   background
 particulate is less  than  1 percent  of  the standard, then it  is
 assumed to  be zero and  background samples  need  not be  taken
 with each exhaust sample.

      Analysis   of  Comments  and  Recommendations:   EPA  believes
 that   filtering   dilution   air   or   accounting  for  background


particulate levels  is  good engineering  practice.  However,  if
background  particulate  levels  are very  lew,   there  will  be  a
negligible error  in  the emission  results.   In any event,  any
error   will   only   overstate   true   particulate   emissions.
Therefore, it  is recommended that the manufacturer  be  given the
option to  control  or account  for  background particulate  as  it
sees fit.

     Calculation of  Measured Particulate Mass

     Section 86.1342-87 of the proposed  regulations states "The
mass  of  determined   from   the   following
equation  when   a   heat  exchanger   is   used   (i.e.   no  flow
    ----- -   -----        -   p      p      ...	
     Pmass-  (Vmix + Vsf)  x ( _f  -  _bf ) x  (1 - 1/DF)
                               V      V
                                sf     bf


     Vmix -      Total    dilute    exhaust    volume    (standard

     Vsf »       Total   volume   of  sample   removed   from  the
                 primary tunnel

     Pf -        Mass of particulate on the sample filter

     Pbf -       Net  weight of  particulate  on  the  background
                 particulate filter
  ;  *   '
     Vbf »   ..    Corrected  volume  of   primary  dilution  air
                 sampled by background particulate sampler

     DF -        Dilution factor

     There  are three  issues  here.   They are:   1)   should the
particulate mass on  the filter plus the background be corrected
for  dilution  factor effects,   or  should just  the background be
corrected,  2)  should  the calculation  be  based on Vmix  or the
sum  of  Vmlic  and  V,f,  and  3)   which  equations  should  be
specified  for  systems  other  than  flow  systems  with  a heat

     Summary of Comments:   EMA commented on  all  three  of  these
issues with the  following statement.  "The proposed  equation is
both  in-error  and  is   inconsistent  with   all  the equations
published   in  the   Final  Rule   for  Gaseous   Emissions   (48FR
p.52236)  §86.1342-84(c) .   In  all  the equations  (1)   through (4)
of  this  paragraph,  HC, NOx,   CO,   and  CO2  mass  are  calculated
based  on Vmlx and  not  on the  sum  of  Vm, x  + V$f.    V, f  is


 not   a   significant   portion  of  V.,,, x,   typically  V,f   is   less
 than  0.1 percent  of  Vm, „ and  can  be  ignored  and  it should  te
 just  as it  is  in the  final  rule for gaseous  emissions.  Also,
 all   these   gaseous " equations   correct   only  the  background
 measurement   by  the  dilution   ratio  effect.  Therefore,   the
 equation for Ptnass should be:
           P       P
           _!      bf
   Vmix x  V    -   V    x  (1  -  1/DF)
           sf     bf
	Other    sampling   procedures   will    require    different
 equations,  e.g.,  proportional  mass  flow  control  system  and
 systems   where   only  secondary   dilution   air   is   filtered,
 manufacturers  should have  the option to use  alternate  equations
 compatible  with their systems and  good  engineering  practice."
      Analysis    of
     	     It   is
technically  correct  that  only  background  should be  corrected
for dilution factor affects (this was a typographical error).

     It is  also true that the current  equation  only applies to
certain system  designs.   Thus,  use of other equations  that are
based on  sound engineering principles,  should  be permitted for
alternate  systems,  but  subject   to  prior  approval  with  the
alternate system itself.

     However,  while  V$ f  is  small  for  many systems,  including
essentially  all gaseous pollutant  sampling  systems, with some
double  dilution   particulate   sampling  systems  it  could  be
significant.   Therefore,   Vsf  should  continue   to   be  included
in  the  equation,   if   significant.   However,   little  accuracy
would  be  lost  if  V, r  were  ignored  if  it  was  less   than 0.5
percent of Vmi„.
the  recommendation  is  that   1)   sampling   volume
 (V,r)  be  retained  in  the  equation,   if  it  is   less  than  0.5
 percent  of Vraix,  2) only  background  be corrected  for  dilution
 factor  effects,   and  -3}  other  equations  be   permitted,   if
 approved in advance by the Administrator.                 -_

      Balance  Requirements

      Section  86.1312-87(b)  of  the proposed  regulation  requires
 that the balance  used  to determine the  weights  of  all  filters
 shall have a  precision and readability of  one microgram.
      Summary of  Comments:   EMA  does not  believe that  the  one
 microgram balance is necessary because the  accuracy  gained  does
 not justify  the  additional expense  and  increased
 associated with  the  one  microgram  balance.    In
 EMA's  knowledge   there  are  not  any one  microgram  electronic
                                       weighing  time
                                       addition,  to


 balances  available that  have  weighing  chambers large enough  for
 the  90  TOT or 110 mm filters  that  are  used on the  EPA  transient
 test  cycle.

      EMA  also presented  the   results  of  an  analysis  that  was
 conducted  that  compared the  overall  accuracies expected with  1
 and   10  microgram  balances.    The  10  microgram   balance  was
 analyzed  assuming a precision  of 20  micrograms.   EMA  concluded
 that  although  the  1   microgram  balance   improves  the   filter
 weighing  accuracy by a factor  20,  this  accuracy is  lost  in  the
 particulate  equation where  other measurements are  included that
 have  1  percent,   2 percent, or  even 3  percent uncertainty.   The
 net  effect  is that the 1 microgram balance,  as compared to  "the
 10  microgram balance with a  precision of  20  micrograms,  reduces
 the   error  by only  .02  percent,  from  5.30  percent  to  5.28
 percent.   (This  was  calculated with a filter loading of  4  mg.)
 EMA  believes that this example  illustrates  the  fact that  there
 is  little benefit in  having  one measurement  substantially more
 accurate  than other measurements used  in the  same  process.

      EMA  also makes  an argument  about the cost of  balances.   A
 typical  10  microgram balance costs approximately  $3,000  but a  1
 microgram balance costs approximately $7,000.  They  feel  that
 the  additional  expense  of  a  1  microgram  balance  should  not be
 forced  upon manufacturers  because the  above  analysis  does  not
 justify  it  in terms of  gained accuracy.

      Analysis of Comments  and  Recommendation;   EMA's  analysis
 of  errors  contained  in their  test procedure comments  appears
"fundamentally sound.   The affect of  using  a  balance  with  a
 pre'cision of 20  micrograms  and  a  readability of  10 micrograms
 appears   minor   and  thus  it   is  recommended  chat  the  test
 procedures  be changed  to reflect this.

      Filter  Reweighing

      Section 86.1339-87  of  the  proposed  regulations  requires
. that  if a  filter is removed  from the weighing chamber  and not
 used within one  hour,  it must be reweighed.

      Summary of   Comments:  EMA sees   no justification for this
 requirement   and  recommends   its  deletion.    They  argue  that
 "there  can  be occurrences  when an unscheduled test delay occurs
 and  filter  and   holder assemblies remain out  of  the weighing
 chamber  for more than  one  hour.   During this delay,  the  filter
 disc may  be installed in  the sealed holder  and   no changes in
 dust or moisture content  could  occur.  If the filter assembly
 was   installed in the  test  fixture during  this delay and some
 moisture penetration and deposition could occur,   more moisture
 deposition  will  occur   during  subsequent sampling  of the exhaust
 gas mixture. All moisture  deposition  either prior to  or during


sampling that  is condensed on  the  filter will  become  adjusted
co the weighing  room's  moisture level during  the stabilization
period prior to  final weighing."

     Analysis  of (Torments  and  Recommendation:   The  purpose of
the  rules  regarding  reweighing is  to  reduce  water  vapor  and
particulate  contamination  of   filters  from  sources  other  than
test-generated   exhaust.    If   a   filter   is   installed   in  a
completely  sealed filter  assembly,  or a  sealed  filter  holder
assembly is  placed  in the sampling line through  which  there is
no flow,  then such  contamination  should  be so  negligible  that
filters should be able  to  go  up to  8  hours  before  they would
have to  be  reweighed.   However,  if  these conditions  of  filter
placement are  not met, then filters  should be  reweighed after 1
hour.  Thus,  it  is  recommended  that  the requirements be changed
to:  1) specify  reweighing  after  8 hours if the  filter is  in a
sealed  holder assembly  or  in  a  sealed  assembly mounted  in  a
sampling system  through  which  there  is no flow,  and  2) specify
reweighing   after  one   hour   if   the  -above  filter  placement
criteria are not met.                        ...

     "Sandwich"  Filter Handling and Weighing

     Section  86.1339-87  of the  proposed rules   requires  that
both the  primary and backup filter  be  weighed independently so
that the  ratio  of   their  net  weights  can  be  determined.   The
backup  filter  net   weight  is  deleted  if   it is  less  than  5
percent of the total.

     Summary  of  Comments:   EMA  comments that  "Some  EMA members
weigh  both  primary   and  back-up   filters  together  as a  pair.
Then, after  sampling, in  removing  filters from the holders, the
back-up filter is inverted  on  top of the primary  filter placing
both  faces   with  sample  accumulation  'sandwiched*   to   the
inside.   This procedure  reduces  the  potential  of  lost  sample
since  now  the  filter   'sandwich*  can be   handled  with  tongs
anywhere  including   the  center.   This  is  especially desirable
with large diameter  filters which tend  to sag  when supported at
the  end.   Weighing  as a pair will, of course,  reduce the number
of.required  weighings, but  will not  permit  the determination of
the  ratio   of  the   net  weights  which  is   the  manufacturers

     Analysis  of Comments  and  Recommendation;   The  procedure
that EMA discusses  appears to be technically  sound.   Loss of
sample  from  filters  that  are  weighted  individually  does  not
appear  to be  a  problem  at the present,  but  the EMA procedure
appears to reduce the likelihood of sample loss even further.

     The  rule  allowing   a  laboratory  to  not  count  up  to  5
percent  of total particulate filter  loading due  to particulate


on the back-up filter is another point that  EMA  brought  up that
also  deserves  analysis.  whereas  this has  been a part  cf  the
HDD particulate testing procedures  from  their inception,  it  is
not good practice  since it  allows  up to a 5 percent error which
could easily be  avoided.   This  change will  not be  mace during
this  rulemaking  because  prior  notice  has  not  been  given  and
some may consider  it  an increase in  stringency.  Nevertheless,
its elimination should be considered in the future.

     The   recommended   action  on   this   issue   is   that  the
"sandwich" filter handling and weighing procedure be permitted.

     Provision for Automatic Data Collection Systems

     Section   86.1310-87(b)(5)(iii)   of   the   proposed  rules
specifies  that  "Chart  deflections   should  be  converted  to
concentration   before    flow    compensation   and   integration"
(underlining added).      	

     Summary  of  Comments;    Ford  feels  that  this  does  not
account  for   automatic  data  collection   (ADC)   systems  and
therefore,  should  be changed to  include  chart  deflections  an'd
analyzer voltage output.

     Analysis  of  Comments  and Recommendation:  This   section
dates from a period when ADC  systems were  generally not used.
ADC systems  are  now common  and therefore  Ford's  recommendation
is  quite  reasonable.   Therefore, it  is  recommended that use of
analyzer voltage output be permitted.

     Hot-Start Restart  for Reasons Other Than Engine Stall

     The current  regulations  for  gaseous  emissions  (Subpart N,
Section  86.1336-84(c)(3))  provides  for  a  hot-start  restart if
the  engine  stalls,  but  no  provision  is  made  for  hot-start
restart  after  operator  error  or  other  small  malfunctions that
can void a test.

     Summary of  Comments:   Caterpillar suggested  including test
voiding  in the wording- for hot start  restarts.   They feel that
this would improve testing efficiency.
     Analysis   of   Comments  and   Recommendation:    The   rules
 regarding  hot-start  cycle  restarts  were  revised  to   include
 equipment  malfunctions   and  were   published  in  the   Federal
 Register  as  technical  amendments  on December  10,  1984.   These
 changes  should  adequately  address  Caterpillar's  concerns   on
 this issue.

 II.  Recommendations Not  Accepted by  EPA

     Six  comments  addressed aspects of the  procedure which have
 the potential  to  substantially  affect measured particulate.   In


 no  case was  there a substantial  amount  of data  available  upon
 which  to base a  decision.   However,  in every case  the  available
 data  indicated that measured particulate mass  could  be affected
 and   thus   that   the  current   specification  was   necessary  to
 prevent  biased measurements  and  unnecessary variability.   In  a
 few   cases,   the  analysis  indicated  that  even  the  present
 specifications  may  allow  undue  variability  in  particulate
 measurements.   These  aspects  of   the   procedure  should   be
 reevaluated in the near future.

      Location of Sample Line  Temperature  Specifications

      Section  86.1310-87(b)(1)(i)(A)  of  the proposed  regulation
 specifies  a  maximum temperature  of  125°F  at  the  sampling  zone
 (in  the primary  tunnel)  for  single-dilution  systems, but  for
 double dilution  systems,   the  125°F  criteria  applies  at   the
 filter face.

_ ._   Summary of Comments:  EMA recommends that the  requirement
 to  be below  125°F  at  the sample  zone  for  single dilution  be
 changed to refer  to  125°F or  less at the particulate filter,  in
 line  with  the double-dilution temperature  requirement.   The EMA
 feels  that  this  temperature  limit is generic in  nature  and.not
 dependent   on the  type of  sampling  system  used;  i.e.,  this
 temperature  limit  and location  should  also  apply   to  single
 dilution systems.

      EMA  also presented data that  they  feel  indicates  that  the
 sample zone  temperature has no  influence on  the  single-dilution
 particulate  results.   This  data  compares  simultaneous  samples
 taken  with  a  single-dilution  system   and   a  double-dilution
 system.   The single-dilution  system had  a  peak  sample  zone
 temperatures in the  220°F range yet  peak filter  temperatures of
 about 110°F.   The  heat  loss  was.  taking  place   in  the  sample
 transfer  tubing  and filter holder.   The average  difference  in
 particulate mass  results between  the  two systems  was  less  than
 0.5  percent.

      In the  EMA-EPA  meeting  of  January  28,  1985,   it  became
 apparent   that the  main  issue here was  the  amount  of  heat
 transfer  that can be permitted in the sample  transfer sections
 of  the single-dilution or, for that  matter,  the  double-dilution

      Analysis  of   Comments  and  Recommendation:    EPA's  diesel
 particulate sampling system specifications are based  on several
 precepts,   two  of which  relate  to  the  issue raised by  EMA.
 These are:   1)  exhaust  should  be  cooled to 125°F  or  less prior
 to  particulate  sampling,   and  2)  this  should be done  to  the
 greatest  extent  possible  by  convection  (i.e.,   using  dilution
 air)   as  this  is  the  manner  in  which   exhaust  from  an  in-use


 engine  is  cooled  in the atmosphere.  The  issue  here is not  the
 125°F maximum temperature but rather how to achieve  it.

     The   criteria   for  heavy-duty  single  dilution   sampling
 systems  came   from  those  for   light-duty   (LD)   particulate
 sampling  systems,  which  is  the  area  where  most  of  the  data
 exist  with  respect  to  testing   procedures.   The  light-duty
 criteria   (which  is  a  single   dilution   system)  is  a maximum
 temperature  of  125°F  or  less  in the  dilution  tunnel.    This
 reflects  EPA's desire  to  maximize heat  transfer  by  convection
 (i.e.,  all cooling  must  take  place in  the  tunnel)  and  limit
 conductive heat  transfer  (i.e.,  heat  loss in  the sample  line
 cannot be  used  to reach the  125°F  limit).

     For heavy-duty (HD)  particulate sampling,  the same  tunnel
 maximum  temperature  of  125°F  for  a  single   dilution  system
 represents a direct  extrapolation from  LD experience and  is,
 technically,  the most  desirable system.   However,  for KD  this
 requires very  large CVS systems (and  large costs)  and thus  EPA
 has  allowed  the   alternate,  double  dilution  system.    EPA's
 intent  for this double  dilution system  is the  same as for  the
 single  dilution  system;   to  achieve  the majority of cooling
 through  convection.    In  establishing  the  specification .for
 temperature  (125°F)  for  this  double  dilution system,   it  was
 applied  to the filter  face  rather  than  the tunnel  since  all  of
 the  tunnel flow is filtered and the  end of the dilution  tunnel
 is  essentially the  same as  the  filter face  (i.e.,  it does  not
 matter which is specified).

     The  data  presented  by EMA  (see  Table A-l)   compares  the
 particulate  results from  a  single dilution system  experiencing
"a minimum  of   110°F  of  conductive cooling  to  results  from  a
 double   dilution  system  which   also   appears  to   allow  much
 conductive cooling.  The  double  dilution system used  conforms
 to  EPA  regulations,  the specifications  for which were  made with
 two  purposes  in mind.  One was   to  limit conductive  cooling.
 The  other  was  to allow  reasonable lengths  of  transfer  lines,
 etc.,  for  ease of  assembly and  location  in the test  cell.   It
 appears  that the flexibility  granted  may  have  been  excessive,
 as. it was not  the  intent  of EPA to permit excessive conductive
 cooling  from the  double dilution system.   Thus,  at  issue  is not
 so  much the  single  dilution  system  specifications but  rather
 those  of  the   double  dilution  system,  which  may  have  to  be
 modified in order  to reduce the allowable amount  of  conductive

     While there is  a  limited  amount of  data  which  shows  the
 effect  of  conductive  heat  loss from  sample  transfer  lines  on
 particulate  concentrations  (particulate increases  as  the  degree
 of   conductive   cooling   increases),[1]*  what  is   available
      Numbers in brackets  refer to References  found at  the end
      of this section.

                                   Table A-l

            Simultaneous Particulate Sampling - EPA Transient Test

Dilution System
Peak "Tenp.'F
(Hot /Cold) Sample
. 21
A -
- H
B -
""' " 223
, „ :, 213
6 Cyl.,
_ 109
Double-Dilution System
D.I. Diesel
' .591
' .629
D.I. Diesel



- 80

                                                                          %  Dif*


                                                                          "  1.6
*  Percent difference, single dilution compared to double dilution.


indicates  that  conductive  cooling  should  be  limited  to  the
fullest  extent  possible.   None   of   it   argues  for  further
relaxations.   Thus,  it  is recommended  that  no changes  be  made
in the  specifications for  single  dilution systems,  since  this
system  still   represents   that   which   is   technically   most
desirable.   No  tightening  of the  specifications  for  the double
dilution system  should  be made  at this time,  since  none  were
proposed.   However,  the  specifications for  the  double dilution
system should be reevaluated  in the future to determine  if  the
degree of conductive cooling currently allowed is acceptable.

     Sample Flow Specifications  and Proportionality

    .Section   86.1310-87(b)(6),   paragraphs   (1)(B   and   c),
(ii)(E)(l and 2) and (ii)(G and  H) ,  require that  the  gas stream
temperature   into    the    particuiate   sampling   system   flow
instrumentation and sample pumps be  maintained at  77° + 9°F,  and
also  that  certain temperatures  be maintained within limits of

   -  The intent  of  these  proposed  requirements  is   to  assure
accurate measurement  of  both the  exhaust  sample mass extracted
from the primary tunnel  and the mass of the  secondary dilution
air  entering  the particuiate system.   This allows establishing
a means for maintaining the  proportionality between the  primary
tunnel mass flow and the extracted exhaust sample.

     Summary  of  Comments:   EMA  and  GM  expressed  in  their
written comments  that they believe  that  this  section  of  the
regulations should provide  system performance requirements,  but
should  not mandate  the  means   by  which   such  performance  is
accomplished.        --           -         	

     In the EPA/EMA  meeting  subsequent to the submission of the
written comments,  it became apparent  that an additional  major
issue of concern  is  the  issue of proportionality between tunnel
and  sample flow.   EMA's  position  is  that:    1)   the proposed
regulations  currently  permit  a  ^5  percent  deterioration  in
sample  flow  from   the  set  point  for   non-flow   compensated
systems, and  this  same  + 5  percent tolerance should be permitted
for  flow compensated systems,  and  2)  a +2 percent  flow change
specification is  permitted for the  main  tunnel  flow, and these
two  flows  (tunnel  and sample   lines)  are  independent and   thus
the  permissible  limits  should be added to  permit  a total of ±7
percent deviation from proportionality.

     Analysis  of  Comments  and  Recommendation;   The proposed
temperature  requirements  for particuiate  sampling  system   flow
instrumentation  and  sample  pumps  are  appropriate  for   some
systems  but  may  not be  appropriate for  others.   This  can be
addressed  by retaining  the current  proposals for  sample   flow


handling  and measurement  but adding  a  provision  that  permits
alternate  systems  if  these  are  shown  to  yield  equivalent
results  and  if  approved  in  advanced  by  the  Administrator.
Section  86.1310-87(a)(7)  contains a  similar  statement,   but  it
is not  clear if it pertains  to particulate sample flow handling
and instrumentation systems  and,  thus,  the  above clarification
will be useful.

     The  proposed  rules are  not  adequately  clear on the limits
of  proportionality.    The  rules  should  be  made explicit  and
uniform   for   both  types   of  systems  (flow  compensated  and
non-flow  compensated).    The  question   is  what   should  the
specifications  be. •

     EMA  believes that tunnel  flow  and  sample  line flow are
independent  and therefore the  tunnel  flow limits  (±2  percent)
and  nonproportional  flow  limits  (+.5  percent)  should  simply be
added  together to  yield   overall  proportionality limits  of  ±7
percent.   While  the   independence  of  these  two  errors  can  be
debated,  the  issue  here  is  not  equity,   but   accuracy.   The
errors  allowed for the currently specified  system were derived
from  the limits of equipment,  not a  decision  that  the errors
were   the   lowest   desirable.     Flow-compensating  equipment
available commercially is  capable of meeting  a  +5 percent error
specification    at    a   reasonable    cost.   "~ The    overall
proportionality  limit  of   flow-compensated   systems   should,
therefore,  remain at   the ±5  percent  level  contained   in the
proposed  rules.  However,  this  level of  non-proportionality (±5
percent)  may  itself  be  excessive  and  should  be  studied
further.  EMA  has  stated  that  they will be submitting  data on
this issue, which  should be useful for this purpose.

     Thus,  in  summary  the recommended resolution of this  issue
is  that 1)   the proposed   flow handling  and  measurement  wording
be  retained,  2)   a  provision be  added  that  permits  alternate
systems  if  these are   shown to yield equivalent  results,  and 3)
a  clarification  be  added  which  states  that  the ±5  percent
proportionality limit  applies  to  both  flow  compensated  and
non-flow compensated systems.

     Test Cell  Temperatures During Natural Cooldown

     Section   86.1334-84   requires  the  test  cell  temperature
during  natural  cooldown to be 68  to 86°F.

    • Summary  of  Comments:   EMA   states  in   their   written
submission  that:

      "...None  of  the  engine manufacturers  have  the capability
     of  cooling  the   test cells  to  assure  that   the  natural
     cooldown  temperature  limit  can  be  met.  If  the  limit can


     not be met,  then a test may  be  postponed  until  the weather
     cnanges.   This  practice  is  currently inefficient,  but  it
     will   become    intoleraole   when   Selective   Enforcement
     Auditing  becomes effective."

     "In Section 86.1330-84  the  cell ambient  temperature during
     the transient  test  is  not  required  to  be  controlled  for
     engines which  do  not have  temperature  dependent auxiliary
     emission  control  devices.   The logic  used  for  the  cell
     ambient temperature  during  the transient  test  should also
     be applied to the natural cooldown."

     In  an EPA/EMA  meeting  subsequent  to  the  submission" of
EMA's comments,  this issue  was  discussed further.   One aspect
of  the  discussion  centered  on  the fact that  two  different
temperatures are  specified  in  the  Code of Federal  Regulations
(CFR) at which a  cold  start emissions  test can  be  started.   If
the engine  is  force  cooled,  it cannot be  started unless the oil
sump is at  75°F,  yet if  the engine  is  .naturally cooled  it  can
be  started  at  86°F.  This  requirement  has been  in  place since

     Analysis  of  Comments and Recommendation;   The  fundamental
purpose for cooling  an engine  by either natural or forced means
is to bring it to a  temperature  that is somewhat representative
of   an   in-use   engines   cold   start.    This   is  particularly
important for the measurement  of HC and  particulate emissions,
since emissions  of  these pollutants tend to decrease as cold
start  temperatures  increase.   The  temperature  specification
that  was  selected   for   natural  cool  .down  was  77°F,  with  a
tolerance  range  of  +9°F.   This  was  based  on  the  current
light-duty practice.  Ever,  though the  upper limit of this range
is 86°F, good engineering practice would  dictate a target value
for  natural cool down of  778F,   and this  is  in  fact the intent
of the rule.  The fairly wide tolerance band is  due  to the fact
that most  test  cells due not  have  precise temperature control,
particularly in the summer.

     A  forced  cool  down  procedure  was  added  at manufacturers'
request  to  shorten  the time necessary  to  prepare an engine for
a  cold-start  test.   The  upper  temperature  limit  of  75°F  for
forced  cool  down  is  consistent with the  natural  cool  down
procedure  for  two  reasons.   Since it is  relatively  easy  to
control  the  final temperature of a forced cool  down, there is
no  need to specify a wide tolerance  band  about  the desired
target.   There  is   no  practical difference  between  77°F  and
75°F, particularly considering that  the forced cool  down occurs
much quicker than the  natural  cool  down and,  thus, some  rebound
in  temperature  is  likely to  occur.  Also,  since  forced  cool
downs  are  performed  to  save  time,  it  is  reasonable to expect
that they  will  be stopped  as  soon  as  the required temperature


is  reached.   Thus,  if 86°F  were the  upper  limit,  this  would
also be  the  average.   The sane  should  not be  true  for natural
cool downs since  manufacturers"  are  not expected  to purposely
control  the  overnight temperatures  of  their  test  cells  to the
upper-limit  86°F  temperature.   Thus,  unless data  are supplied
demonstrating  that  higher  cold  start  temperatures  have  no
effect on  emissions,   it  is  recommended  that  no changes be made
in the cool down procedure.

     Practically   speaking,    rejecting   EMA's   recommendation
should only  have  a minor  economic  impact  on  test costs.   While
air conditioning  test cells  to ensure  temperatures  below 86°F
for natural  cool downs can be  quite  expensive,  this is not "the
only alternative  available to  manufacturers.  The  forced cool
down  procedure  can be  used.   Some manufacturers  objected to
'this,  due  to  the  need  to use  city water   to  reach  the 75°F
limit.   However,   internal cooling water can  be used to provide
most of  the  necessary cooling and the cooler city  water  can be
used   to  provide  the   last   10-20°F  of   cooling.   While
constituting  some  cost,  the overall  cost  is  less  than that of
the water  itself,  since this water will be added to  the cooling
water system within the lab and  recycled.

     Dilution Air  Temperature Limits

     Sections    86.1310-87

     "Due   to   the  leadtine   necessary   to   construct   these
     facilities,   in   anticipation   of   the  gaseous  FTP,  the
     manufacturers  were  led  to  believe  that  the  CVS  systems
     constructed  to meet  these  procedures  would  also  suffice
     for   the   impending   particulate   test   procedures.    To
     redesign  and  modify these established  systems in order to
     add   the   necessary  cooling   capabilities  would   be  a
     difficult  and  expensive  task  for  the  manufacturers  and
     could   possibly   force  the   relocation   of   entire  CVS
     systems.   An  industry  estimate  ranging  from  $280,000 to
     $420,000  has  been obtained  to  equip  test cells  with the
     necessary cooling capacity and controls."

     The  EMA  is   in  support of  the 125°F  maximum temperature
     requirement   at   the   particulate   filter  holder.     This
     temperature   limit  effectively  necessitates   primary  and
     secondary  dilution   air temperatures  to  be   significantly
     below   125°F.    In   essence,   the    particulate   filter
  ~~ "temperature  requirement indirectly regulates  the dilution
     air  temperatures  to  practical  ranges.   As  suggested in SAE
     Paper  800185,[2]  little  is  known  about  the  influence of
     dilution  air  temperature on  particulate measurements since
     investigations to date  have  not separated  the  dilution,air
     temperature   factor   from   other   dilution  and  sampling
     effects.  What can be said is  that the combined effects of
    , many   of  these  factors  on  particulate  measurements  are
     small,  in  the range of ambient to  125°  F, suggesting  that
     any  variations  in  dilution  air   temperature would   have
     insignificant effects on particulate measurements."

     "The  EMA  recommends  that  the dilution  air  temperature
     range   (68  to  86°F)   requirement  be   modified  to   allow
     temperatures  above  86°F,  provided  the  dilution air  is not
     artificially  heated  above  this  temperature.   This  would
     save  the manufacturers  the cost of adding  cooling capacity
     to  their  dilution air systems   in  order  to provide for  high
     ambient temperatures occurring  during warm summer months."

'-• •  Analysis   of  Comments   and   Recommendations:    EMA  has
 suggested   that   little  is  known  regarding  the   influence of
 dilution  air  temperatures  on  particulate  mass concentrations.
 While  this  is  partially true,  there  are  some data  that  show
 that  dilution  air  temperature  is   potentially  a  significant
 factor.   These data are  presented  by  Reichel  et  al.,[l]  where
 they •  show  a  23   to  33   percent  decrease  in  particulate
 concentration  when  the  dilution air  temperature  is  increased
 from  68°F  to   122°F.   When  the  dilution  air temperature is
 increased    from   86°F   to   122°F,   the   tunnel  particulate
 concentration  decreases  by about 17  percent.   These  reductions
 in   particulate   concentration  are   primarily   due   to  the
 desorption  of  organics,   according  to  the  authors' theoretical


calculations  and  thermogravimetric  observations.   EMA  did  not
specify  how  much  in  excess  of-85°F they  would like  the  upper
limit  for  dilution  air  temperatures  and  the  dilution  air
temperatures  in the  manufacturers'  facilities  would  not likely
reach  the 122°F  of  the  above cited  data.   Nevertheless,  the
data   do  indicate   a   significant  effect   on   particulate
concentration due to dilution air temperatures.

     EMA  also refers  to  EPA's promulgation  of the  final  rule
for heavy-duty gaseous  emissions  and they imply that  this  rule
also  included  all   of  the  provisions  needed for  particulate
measurement.   Actually,  numerous  changes  in  the  final  gaseous
test  procedure  requirements  were  made  by   EPA  so  that  the
manufacturers would  not  have to invest in  equipment  needed for
particulate measurement at  that time,  if not so desired.  EPA's
intent  in so doing  was  to delay  particulate  testing equipment
requirements  such  that there  could  be a  more  ordered  phase-in
for  these equipment needs.   One  way  of doing  this  was  to
•attempt  to assure  that new  equipment .that would  be purchased
for compliance  with the  gaseous  emissions testing  rules  would
also  be  useful when  particulate  testing was required.   For
example,  use  of a dilution tunnel  was  allowed  under  the gaseous
emission   regulations,  but  was  not  required.   However, .the
gaseous   rules  and  supporting documents   did  not  imply  that
particulate  testing  would not  require additional  equipment and
specifications  such  as  secondary  dilution   tunnels,   weighing
balances,   and   dilution   air  temperatures.   An    additional
observation  on  dilution  air  temperature  limits  is  that  these
limits  have  also  been  in effect  for  seven years  of  light-duty
particulate testing.

     Therefore,  since  dilution  air  temperatures  can  have   a
substantial  effect  on particulate  emissions,  no  changes in the
proposed  dilution  air temperature  requirements should  be  made
until  some further  date  when sufficient  data  are  available to
establish  that  no   effect   is   present   or   to   establish
satisfactory  correction factors.

     Humidity  Effect   Correction    Factor    for   Particulate
    'No  humidity-related  correction factor currently  exists for
particulate measurement.

     Summary  of Comments:   EMA submitted  a   limited  amount of
data on  the effects  of  humidity on particulate measurements and
intends  to submit additional  engine data  at  a later  date.  The
current   set  of  data  show   that   the effect  of  humidity  on
particulate   is    in   the    opposite  direction   and   about
three-fourths the  size of that for  NOx.   The  equation  would be
of  the same  general  form  as  the  NOx humidity  correction factor


     EMA  recommends  that   EPA   consider  a   humidity  effect
correction  factor   for   particulate  measurements   using  the
submitted data,  with the option of  accepting  additional data at
a later date.

     Analysis  of  Comments  and Recommendation:   The  data that
are available on this subject  are very limited  (see  Table A-2)
and  are  an  inadequate  base  upon which  to  formulate  a  rule
change  of the  magnitude  suggested  in   the  EMA  comment.   In
particular,  no  data  exist  on the  impact  of   various control
technologies  (e.g., trap-oxidizers) on this  effect.   Therefore,
it is  recommended  that   resolution  of  this issue  await receipt
of additional  data.

     Sulfur Correction Factor

     EMA  suggests  that   a  sulfur  correction  factor  similar  to
the NOx  humidity  correction factor  be  employed to  correct for
the observed  increase in  particulate with an  increase in fuel

     Summary of  Comments:   EMA  cites data  that  show that for
each  0.05  percent  fuel  sulfur  mass  increase,  there   is   a
corresponding  increase   in  measured  particulate  emissions  of
0.024  g/BHP-hr  due  only to  the   change  in  fuel  sulfur.   EMA
suggests  that  a  correction  factor  be used to  correct for this
perceived  inequity.   Furthermore,  EMA  believes  that  a   sulfur
correction  factor   will  become  more  important as  particulate
standards become more stringent in the future.

     Caterpillar raised  a similar  concern about  the  inclusion
/if  uat-or  /uhinh  i <;  a 5<;nr*i al-«arl  UJT hh  «n1fat-
                           Table A-2

                     Calculated "A" Values
          For  Particulate  Humidity Correction Factor
Cummins #1
Cummins #2
Mack *1
Mack #2
Mack *3
Mack S4
Mack #5 __J._.
Mack *6
Mack #7
Mack #8
Mack #9
Mack 676
IHC #1
Cummins 903
DDAD 871
Caterpillar fl
Caterpillar 12
Caterpillar #3
Caterpillar £4
Caterpillar #5
Caterpillar *6
Calculated "A"*
+.00099 ~
+.00303 . . ..
Avg. 0.00170
     Equation  for Correction  Factor:
                                                      Mean Part.

                                                          .49   --•
Corrected Particulate  =  (
                          1  +  A  (Humidxty  -  75)
)  X Observed Pars, icu i ^ ce


 +0.05   weight  percent   sulfur  or   less.    Taking   Chevron's
 relationship  at  face value,  this change  in  sulfur levels  could
 result  in  a  change  in particulate  emission  levels  of  +0.024
 g/BHP-hr,  which  is  +4  percent  of  the  0.6 g/BHP-hr particulate
 standard.   While  this  effect would represent a greater  percent
 of  a 0.25  g/BHP-hr  particulate standard,  use  of particulate
 control  devices  such  as  traps  should  reduce  the  size  of  the
 fuel  sulfur  effect  somewhat.    Nevertheless,   this  degree   of
 potential variability  is larger  than  generally  desired.

     The   relatively  wide  specification  for  sulfur   content
 allows  the  sulfur  content  of the test  fuel  to  change with  that
 of  commercial fuel  without  requiring modifications to the  CFR,
 which  are  costly  and  time consuming.  This   flexibility  is
 intended,  from EPA's  point  of  view,  and should  be maintained.
 Use of  a  sulfur  correction factor  would  necessarily  require
 that  a   target  fuel   sulfur  level  be  specified, essentially
 removing  this  flexibility.   As  the  sulfur  content   of  EPA's
 current   (or   projected  future)   test   fuel   is   not markedly
 different  that  that  used to  develop  all  of  the particulate
 emission  data  used  in the  technical  feasibility analysis  in
 Chapter  2,  retention of the current  provisions does  not  affect
 the feasibility of  the standards  being  promulgated as  long  a-s
 the  sulfur  levels   of   commercial  fuels  do   not   increase
 dramatically  in the future.

     The  issue of  in-use  sulfur levels  is addressed  in  Chapter
 2,  as a  number  of  manufacturers  requested  that   in-use  sulfur
 levels  be  controlled  to  lower  levels  by EPA  to allow  use  of
 various   aftertreatment technology.   There  it  was  determined
• that  the feasibility of the final  particulate  standards  was not
 contingent  upon this  control.   However,  it  was  also  indicated
 that  the  control  of  commercial  fuel  sulfur  content would  be
.further  investigated  in the future  as   a  means  of controlling
 particulate  emissions.   Investigation   of   the  potential  for
 in-use  sulfur levels increasing in  the  future  is  a natural part
 of such  a  study.   Thus,  any  potential for  high  in-use sulfur
 levels,  and  thus,   high  certification  fuel  sulfur   levels,  to
 cause   the   particulate standards   to  be  infeasible  will  be
 investigated  at  that  time.   In the  meantime,  with  relatively
 constant   fuel   sulfur  levels,   feasibility  should  not  be  an
 issue.   Thus, it is  recommended  that no changes  be made  to  the
 test procedures to account  for  the  sulfur  content of the test

      Caterpillar  suggested the  elimination  of  the inclusion  of
 water   associated   with  sulfate  in  the  measured  particulate
 mass.   How this  could be done is  not  clear  at  this  point  and
 requires further  study.   However,  ammoniation of the filtered
 particulate is one  possible approach.   As discussed  above,  the
 standards  being  promulgated  are  based  on  measurements  which


 include  such  water.   Removing\-the;CwaterRnow would either  reduce
 the stringency of  the standards or  require  that the  standards
 be  modified.   Thus,  no  action should be  taken  with respect  to
 water  measurement  at  this time.  Further study may  be  merited,
 however,  if  future  sulfur   levels  increase*  or  if  desirable
 future control technology is  found to  affect  sulfate,  and thus,
 water  levels.  This  study should be coupled  with the  analysis
 of  future  commercial  fuel   sulfur  levels and  their  potential
 control  described above.

 III. Issues Raised by EPA in  NPRM

     The NPRM  requested  comment  on  four  issues   because   of
 potential  improvements  were  believed to  exist  in these  areas.
 These  areas  were:    1)  the   possibility  of relaxing  the  cycle
 performance   statistics  of  horsepower   standard error,  2)  the
 possibility  of changing  the  primary torque  measurement  method
 to  an  electronically compensated case   load  system,  3) the  NOx
 correction  factor for humidity,  focusing on the-adequacy  of  the
 current  factor  for  low  NOx  engines and 4)  the  addition of  a
 standard  calibration  procedure   for   HDGE   throttle   control

     In  general, EPA received little response  to these issues.
 From the comments  that  were  received it  can  be concluded  that
 there  is no  dissatisfaction  or known problems  with  the current
 system.   The  only area that  did result  in the  receipt of  data
 was the  NOx  correction  factor,   where the   data  presented
 indicated   that   the   current  NOx   correction   factor   was
.appropriate  for low  NOx engines  as well as current  engines (see
 Tables A-3 and  A-4) .  Thus,  as a  result of  the  comments  and
 analysis of  these four  issues, no  changes  should  be made  in
 these  areas.

 IV. Other  Issues

     The last group of issues do not directly relate to Subpart
 N  but  will  nevertheless be  addressed here  because they  deal
 with test  procedures.

     Smoke  Standards                                       ^--

     EPA  did  not  propose  to  eliminate  the current   smoke
 standards  when it proposed to add particulate  standards.

     Summary  of  Comments:  Mack commented that an  engine  that
 meets  the  0.60  g/BHP-hr  standard will  easily   pass  the  smoke
 standards  and  therefore,  the  smoke  standards  are  not  needed.
 They present  no data  to support this  but state  that  it  is based
 on limited data.


                           Table A-3

                   Calculated "A" Values For
               NOx Humidity Correction Factor —
      Engines  With NOx Emissions  Greater  than 6.0 g/BHP-hr

                          Mean NOx
    Engine               (g/BHP-hr)             Calculated  "A"*

Previously Submitted  Data**

Caterpillar fl               6.68                      -.0017
Caterpillar 42               8.96                      -.0025
Cummins 42                   7.66                      -.0023
Cummins *3                   6.36                      -.0017
Mack  #1                      7.85                      -.0032
Mack  #2                      9.39                     -+.0028
Mack  #3                                             .   -.0028
Mack  *4                      7.74                      -.0037
Mack  *5           •           7.44                      -.0025
Mack  #6                      7.02                      -.0024
DDAD  #3                      6.12            '          -.0029
Mack  676                     7.47                      -.0022
DDAD  871                     7.66                      -.0025

Additional Data

Caterpillar *i               9.11                _      -=0027
Caterpillar #2               7.98                      -.0029

                                          Average A =   -.00259
*    Equation  for Correction  Factor:

Corrected Particulate  =(—-——7-5	r-r-.	TT\)X Observed Particulars
                        L-  +• A (Humidity — /a;

**   Public   Docket   No.   A-80-18,   Parciculate  Regulations   for
     H.D.D.E.,  "Statement  of  the E4A."  September  13,  1982,  Appendix
     "D", p. 2,  Table  i.


      Analysis  o£  Comments and  Recommendation:   Low  particulate
  emission  standards  may lcv;er srroke levels on  average,  but will
  not   necessarily   guarantee   smoke   levels   below  the   smoke
  standard.    This   is   because  the  two   standards  and   their
  associated  test  procedures   are  not  mutually  inclusive  as  to
  intent  and  result.   The purpose  of  the  smoke  standard  is  to
  control  worst-case  smoke levels,  whereas  the  purpose  of  the
  particulate  standards  is to  control  average  transient   cycle
  particulate.    Since   the  engine  operating   conditions   which
  produce  wo.rst-case  smoke are  not  dominant  in  the • transient
  cycle,  a given  engine could  conceivably pass the  particulate
  standard  and  fail  the smoke  standard.  The  implementation  of
  traps  may be  the  one  particulate control  approach that  woruld
  provide  smoke  control,   since  traps  are  effective  under  all
  driving  conditions.   However,  it  is  unlikely  that  all  future
  engines will  be equipped with  traps  and the cost  of  running  a
  smoke  test  is  quite  small.   Thus, it  is  recommended that  the
  smoke  standards  and  their  associated  test   requirements   be

      'OfficialTest  Data

      Paragraph    86.090-29(b)(3)(i)     requires     that     the
  Administrator's data  shall  comprise the official  test data 'for
  any  engine tested.

      Summa ry  of  Comments:   Mack   feels  that  there  is  a wide
  variation  in test results from  facility to  facility with  no  one
  facility   singled  out   as   grossly   superior  or  in   error.
  Accordingly,  in cases  where  the  manufacturer  and  Administrator
-  differ  by more than 10 percent,  Mack  recommends use of a third
  laboratory as  a referee.

      Analysis  of  Comments and  Recommendation:   The  designation
  of  EPA  data  as the official  test  data has  been in effect since
  the  implementation of  emission  standard  in  the  early 1970*s.
  As  no  evidence was presented that demonstrates why  the current
  approach  is  inadequate, it  is  recommended that no  change  should
  be made.

      EPA Approved Equipment

      Paragraph 86.090-29(b)(2)   requires  the   manufacturer   to
  provide   "...instrumentation  and   equipment   specified by .the
  Administrator...."  (underlining added).

      Summary  of  Comments:  Mack  commented  that  "In  the  past,
  manufacturers   have    been    allowed   deviations    from   the
  instrumentation  specified  in  the  Code of  Federal  Regulations
  based on demonstrated  equivalency.   Mack feels that  there  is  no
  reason  to  abolish  this  practice  and the  flexibility that  it


allows.   They  feel  that  demonstrating  equivalency  guarantees
that  the accuracy  of  the  testing will  not  suffer.  Mack feels
that  the wording  should be  changed to  "...instrumentation  and
equipment   approved  by  the  Administrator..."  to  allow  the
manufacturer  the  flexibility to install the instrumentation and
equipment  in  a manner  most suitable  to his operation.

     Analysis of Comments and  Recommendation;   The requirement
for equipment specified by the EPA  has  been in  place  for many
years.   This  requirement  provides EPA with  the  flexibility of
being  able  to specify use  of a particular measurement procedure
or  technique to  enable  a   more confident  assessment  of  the
emissions  of  an  engine.  Whereas  this  flexibility  should be
retained,  it  should be pointed out  that  EPA has  no intention of
being  unreasonable in exercising this provision.   To  date,  EPA
has  rarely, if ever,  exercised this authority  to require use of
special  equipment  with respect  to  heavy-duty  diesel  testing.
Therefore,  it is recommended that this provision be retained in
-its current form.



     1.    "Influence  on Particulates  in Diluted Diesel  Engine
Exhaust   Gas,"   Stefon   Reichel,   Franz   Pischinger,   Gerhard
Lepperhoff, SAE Paper  831333,  1983.

     2.    "Experimental    Measurements   of    the    Independent
Effects  of  Dilution  Ratio  and  Filter  Temperature  on  Diesel
Exhaust  Particulate Samples,"  J.  S. MacDonald,  S.  L.  Plee,  J.
B.  D'Arcy,  and  R.   M.  Schreck,   SAE   Paper  800185,  Detroit,
February 1980.
                                             U S GOVERNMENT PRINTING OFFICE 1985-554-048/10003